Full text of "Journal of the Chemical Society. Transactions 1886: Vol 49"

See other formats

JOURNAL

THE CHEMICAL SOCIETY.

University of difichi

Wl em
Hous

Committee of Publication ;

H. E. Armstrone, Ph.D., F.R.S. | S. U. Proxertne, M.A.

F. R. Japp, M.A., Ph.D., F.R.S. | R. T. Purmperoy,: Ph.D.

R. Metpota, F.R.S. | W. J. Russext, Ph.D., F.R.S.
Hugo Mutter, Ph.D., F.R.S. J. Mittar THOMSON.

F. J. M. Pages, B.Se. T. E. THorpsg, Ph.D., F.R.S.
W. H. Perxrn, Ph.D., F.R.S.

Editor :
C. E. Groves, F.R.S.

Sub- Editor:

A. J. GREENAWAY.

University of 22x23

eo
—

Vol. XLIX.
1886. TRANSACTIONS.

LONDON:
J. VAN VOORST, 1, PATERNOSTER ROW.
1286.

LONDON:

HARRISON AND SONS, PRINTERS IN ORDINARY TO HER MAJESTY, ST. MARTIN’S LANE.

PAPERS READ BEFORE THE CHEMICAL SOCIETY.

PAGE
I.—Modifications of Double Sulphates. Part I. By Spencer
Umrrevitte PickerinG, M.A. aaa Professor of Chemistry

at Bedford College . ‘ ‘ , ‘ , l

1I.—Modifications of Double Salehaten, Part Il. Specific
Heat Determinations. By Spencer Umrrevitte Pickerina,
M.A. Oxon, Professor of Chemistry at Bedford College . 12

III.—An Examination of the Phenol Constituents of Blast-
furnace Tar, obtained by the Alexander and McCosh Pro-
cess at the Gartsherrie Ironworks. Part I. By Watson
Smitx, Lecturer in Chemical Technology in the Victoria
University, Manchester, with Messrs. J. F. H. Courrs and
H. E. Broruers . ; . : , i ; / . FF

1V.—Aluminium Alcohols. Part III. Aluminium Orthocre-
sylate and its Products of Decomposition by Heat. By
J. H. Guapstone, Ph.D., F.R.S., and Aurrep Trise, F.I.C.,

Lecturer on Chemistry i in Dulwich College . . 25
V.—On some New Vanadium Compounds. By J. T. Brrervey,
Dalton Scholar, Owens College . ° 30

VI.—On the Vaponur-pressures of Mercury. By Waseus
Ramsay, Ph.D., and Sroney Youna, D.Sc. . 37

VII.—<Action of Phewher us Pentachloride on Ethy lic Diethyl-
acetoacetate. By J. Wittiam James, enieeied ne of

South Wales, Cardiff . ‘ ‘ 50
VIII.—On the Sugars of some Cereals and of Gemdeateil
Grain. By C. O’Suttivay, F.R.S. . . ' . 8
IX.—On the Presence of “ Raffinose” in Barley 7 C.
- O’Scutiivay, F.R.S. R 70

X.—The annie and Genesis of Sateen at ‘High
Temperatures. I. The Products of the Manufacture of
Gas from Petroleum. By Henry E. Armsrrone and A. K.
MILLER. 74

4 XI.—The Combustion of en Oxide and en By
Harotp Dixon, M.A., the Duke of Bedford’s Lecturer in
Chemistry, Balliol College, Oxford : 4

CONTENTS.

XII —The Theory of the Interaction of Carbon Monoxide,
Water and Oxygen Gases: a Note on Mr. H. B. Dixon’s
Paper on the Action of Carbonic Oxide on Steam. By
Henry E. ARMSTRONG .

XIII.—On the Use of Ferrous Sulphate in Auptodiiom. By
Dr. A. B. Grirritus, F.R.S.E., Lecturer on ne
Technical School, Manchester, &e. ‘

XIV.—On Multiple Sulphates. By Emity Aston ~_ Sue
Umrrevitte Pickerine, M.A. _— Professor of ee
at Bedford College ‘

XV.—tThe Influence of Silicon on the —_— of Cast Iron.
Part III. By Tuomas Turner, Assoc. R.S.M., Demonstrator
of Chemistry, Mason College, Birmingham ;

XV I.—On a Method of Separation and Estimation of Shenson.
By G. H. Batey, D.Sc., Ph.D., Assistant Lecturer on Che-
mistry at the Owens College ‘ , . , ,

XVII.—Notes on an Analysis of Koppite. By G. H Bailey,
D.Sc., Ph.D., Assistant Lecturer on Chemstry at the Owens
College ; . ‘ .

XVIII. a were Acid we some of its Derivatives.
Part IV. By W. H. Perkin mn ), Ph.D., and A. Catan,
Ph.D.

XIX. _The Chemical ‘Action of Pure Caltivations of aint
Aceti. By A. J. Brown

XX.—The Monobromophthalic Acids. By G. einatanis M.A. .

XXI.—Some Derivatives of Thiocarbamide. By Georce
McGowan, Ph.D., F.R.S.E., Demonstrator in ueareie
University College, Bangor .

XXII.—Bacteriological Research from a ‘Biologist’ Point of
View. By E. Kuen, M.D., F.R.S., Lecturer on General
Anatomy and Physiology in the Medical School of St.
Bartholomew’s Hospital, London °

XXTIT.—On the Constitution of Undecylenic Acid, as s indicated
by its Magnetic Rotation, and on the Magnetic Rotation,
&c., of Mono- and Di- allylacetic Acids, and of Ethyl
Diallylmalonate. By W. H. Perkin, Ph.D., F.R.S.

XXIV.—On the Condition of Silicon in Pig a By A. B,
JorpAN (Student in the Chemical Laboratories, Mason
College, Birmingham) and THomas TurNER

XXV.—Reactions supposed to yield Nitroxyl or Nitryl Chloride.
By W. Cottinewooo Wittiams, B.Se., Tangye Scholar in
the Mason Science College, Birmingham

XXVI.—On Sulphine Salts containing the Ethylene Radical,
Part I. Diethylenesulphidemethyl-sulphine Salts. By
Orme Masson, M.A., D.Sc., University of Edinburgh . ;

CONTENTS.

XXVIT.—On Sulphine Salts containing the Ethylene Radical.
Part Il. Dehn’s Reaction between Ethylene Bromide and
Ethyl Sulphide. By Orme Masson, M.A., D.Sc.

XXVIII.—Certain Aromatic aa and Cosbneaiine. By
H. Lioyp Snape, B.Sc.

XXIX.—The Influence of eseiaiens on the Heat of Chemical
Combination. By Spencer U. Pickering, M.A., Professor
of Chemistry at Bedford College.

XXX.—Contributions to the History of —e Chloride al
Cyanuric Acid. By Atrrep Senter, M.D.. ‘

XX XI.—Contributions to a Knowledge of Cyanuric Dateien,
By Harotp H. Fries .

XXXII.—On the Essential Oil of Lime Leaves (Citrus
Iimetta). Preliminary Notice. By Francis Warts, F.C.S.,
Chemist to the Montserrat Company, Montserrat, \V.I.

XXXIII.—The Formation of Acids from Aldehydes by the
Action of Anhydrides and Salts, and the Formation of
Ketones from the Compounds resulting from the Union of

Anhydrides and Salts. By W. H. Perkin, Ph.D., F.R.S. .
Annual General Meeting

XXXIV.—The Relation of Sete Acid am its a
nitro-derivatives. By Cuaries M. Stuart, M.A., Fellow of
St. John’s College, Cambridge

XXXV.—Action of Cinnamic Acid and Salicylic Aldehydes on
Malonic Acid. By Cuartes M. Srvart, M.A., Fellow of St.
Jobn’s College, Cambridge .

XXXVI.—Note on the Action of iii on : Quiet
Dichloride. By Samvuen an B.Sc., eealatated —_—
London

XXX VII.—Some Sulphur a of —— By v. H.
Vexey, M.A., of the Laboratory, Christ Church, Oxford

XXXVIIL—Amidodiphenylsulphonic Acid and Azo-dyes from
Diphenyl. By Tuomas Carnetiey, D.Sc., and James
ScuteseLman, University College, Dundee .

XXXIX.—The Combustion of Cyanogen. By Haroip —
M.A., the Duke of Bedford’s Lecturer in Chemistry, Balliol
College, Oxford .

XL.—The Eurhodines, a New Class of Colouring Matters. By
Orto N. Wirt, Ph.D.

XLI.—Parabenzylphenol wel its Setetiens (Part IID, - on
an Isomeric Benzylphenol. By Epwarp H. Rennin, M.A.
(Sydney), D.Sc. (London), Professor of ena in the
University of Adelaide, S.A. ‘

XLIT.—On Water of Crystallisation. By Sresme Verewmnen
PickErinG, M.A., Professor of Chemistry at Bedford College

vl CONTENTS.

XLIII.—On an Acetic Ferment which forms Cellulose. By
Aprian J. Brown

XLIV.- ~Mieinainelineahiantie Acid wit some of its Deri-
vatives. By W. H. Perkin hin * — and Gustav
Bectenot, Ph.D. ‘

XLV.—On the Tensipanennne of — - i ~
Iodine Monochloride. By Witttam Ramsay, PhD., and
Sypyvey Youne, D.Sc.

XLVI.—On the Action of Aldehydes ‘il poesia on Benzil.
By Francis R. Japp, F.R.S., and W. Parmer Wynne, B.Sc.
(continued) . :

XLVII.—On Imabenzil. By — R. — F.R. 8., ~ w.
Patmer Wynne, B.Sc. .

XLVIII.—On the Action of Pecssiaiie Sulphide on pe
phenone. By Francis R. Japp, F.RS va and JULIUS
RascHEN

XLIX.—A Method for the Setien nt Estimation of
Zirconium. By G. H. Batwey, D.Se., Ph.D., Assistant
Lecturer in the Owens College R

L.—Derivatives of Taurine. Part II. By J. — San
Ph.D., F.C.S., University College of South Wales, Cardiff .

LI.—The Influence of Remelting on the Properties of Cast Iron.
Notes on Sir W. Fairbairn’s 1853 Experiments. By
Tuomas Turner, Assoc. R.S.M., Demonstrator of Chem-
istry, Mason College, Birmingham

LII.—Some Ammonium Compounds and other Derivatives of
a-1’ Hydroxyquinoline. By C. H. Kony, B.Sc., Ph.D.

LIIT.—8-Sulphophthalic Acid. By A. Rés, Ph.D. .

LIV.—Some Compounds obtained by the Aid of 8-Sulpho-
phthalic Acid. By C. Grarse, Ph.D., and A. Réz, Ph.D. .

LV.—Mercury Sulphites and the Constitution of Sulphites.
By Epwarp Divers, M.D., F.R.S., and Tersuxicur Sutwrpzv,
M.E., Imperial Japanese College of Engineering, Tokio

LVI.—On the Electrolysis of Aqueous Solutions of Sulphuric
Acid, with special Reference to the Forms of —
obtained. By Hersert McLeop

LVII.—On Essential Oils. Part III. Their Specific Rofrne.
tive and a a By Dr. J. H. Gt “renee
F.R.S.

LVIII.—A Method of Suniatinabin the Constitution of as
and Diazo-derivatives and Analogous Compounds. By
Rapoaet Mevpota, F.R.S., Professor of Chemistry, and
F. W. Srreatrei.p, Demonstrator of Chemistry in the
Finsbury Techaical a ~~ and Guilds of London
Institute °

591

609

624

CONTENTS.

L1X.—The Formation and Destruction of Nitrates and Nitrites
in Artificial Solutions and in River and Well Waters. By
J. H. M. Munro, D.Sc., a of aieanite Downton,
Salisbury , ‘

LX.—Detection and Estimation of Iodine, Sten. and Chlo-
rine. By M. Decnay, F.C.S., Lecturer on a &e.,
School of Science, Hawick . °

LXI.—Note on the Vapour-densities of Chloral Ethyl- alco-
holate. By Witu1am Ramsay, Ph.D., and Sypney Youne,
D.Se. ' , ‘ ;, . ‘

LXII.—Water of Crystallisation. By W. W. J. Nicot, M.A.,
D.Se., F.R.S.E., Lecturer on ne Mason College,
Birmingham

LXIJII.—Phenylsulphonic Aubpdelie. By J. L. H. " Awpamate

LXIV.—On the Action of Hexabromacetone on Urea. By
ALFRED SENIER . ° .
LXV.—On the Measnrement of the on Forces pro-
duced by the Combination of Cadmium and Iodine in

Presence of Water. By A. P. Lauriz, B.A., B.Sc.

LXVI.—Contributions to our Knowledge of the Chlorides of
Antimony. By Ricnarp Anscniitz and P. Norman Evans .

LXVII.—Pipitzahoic Acid. (First Communication.) By
Ricuarp Anscuititz and Jonn Watrer LEATHER . ; é

LXVIII.—On the Analysis of Alloys and Minerals containing
the Heavy =n eens ns &e. a THOMAS
BayYLey

LXIX.—Further Contributions to the Wiesiiidiiie of —!
Chloride and other a aaceey Derivatives. By Haxotp H.
FRiks . ,

LXX.—Further Chemstinns on the Action of Meniinens
acetone on Urea. By ALFrep SENIER.

LXXI.—The Constitution of Diazobenzeneanilide and its
Relation to Amidoazobenzene. II. By R. J. FRiswe.t
and A. G. Green

LXXII.—Communications hits the iene of Univ exeity
College, Bristol. I. The Estimation of Free Oxygen in
Water. By Miss Karuarine J. Wit.1ams and Professor W.
Ramsay

LXXIII. nadaeamibiailios: of Veowr a of Alcohols _
Organic Acids, and the Relations existing between the
Vapour-pressures of the Alcohols and Organic Acids. By
Artur Ricnarpson, Ph.D., University College, Bristol

LXXIV.—On the Magnetic Rotation of Mixtures of Water with
some of the Acids of the Fatty Series, with Alcohol
and with Sulphuric Acid; and Observations on Water of

Crystallisation. By W. H. Perxiy, Ph.D., F.R.S.

Vii

PAGE

632

682

685

ankle

tT Sar 194:

“rm

CONTENTS,

Vili

LXXV.—Evaporation and Dissociation. Part IV. A Study
of the Thermal Properties of Acetic Acid. By Wittiam
Ramsay, Ph.D., and Sypney Youne, D.Sc.

LXXVI.—Trimethyldiethylamidobenzene. By R. F. —
B.A., M.D., Lecturer on ws in McGill ies:
Montreal

LXXVII.—On the ulin of Secaien : on Phosphoras Tri-
chloride. By A. L. Srery, Scholar in the Chemical
Laboratory of the Mason College, Birmingham .

LXXVIII.—The Electrolytic Preparation of Vanadious Sul-
phate. By J. T. Briertey, Dalton Chemical en
Owens College

LXXTX.—On ‘sneeiin debating of Benzoin. By ee
R. Japp, F.R.S., and W. H. Witson, Ph.D. ‘

LXXX.—Note on a Compound from Benzil and a
Alcohol. By Francis R. Japp, F.R.S., and Jutivs RascHEen

LXXXI.—On Phosphorus Tetroxide. By T. E. Tuorps, F.R.S.,
and A. E. Turron, Associate of the Normal School of
Science, South Kensington

LXXXITI.—A Chemical Study of Vegetable Albinism. Part IL.
Experiments with Quercus rubra. By A. H. Cuurcn, Pro-
fessor of Chemistry in the Royal Academy of Arts

LXXXTII.—Conversion of Ditolane-azotide into Diphenan-
thrylene-azotide. By Francis R. Japp, F.R.S., and Cosmo
Innes Burton, B.Sc.

LXXXIV.—<Action of the ——— on the Salts of i
Bases. Part II. Tetramethylammonium Salts. By
LzonarD Dosstn, Ph.D., University of Edinburgh, and
Orme Masson, M.A., D.Sc., Professor of seid in the
University of Melbourne

LXXXV.—Glycyphyllin, the Sweet Principle of Smilax dian
phylla. By Epwarp H. Rewntz, M.A. (Sydney), D.Sc.
(London), Professor of Chemistry i in the ne of

Adelaide, South Adelaide . ,

790

813

839

843

857

JOURNAL SR

THE CHEMICAL SOCIETY.

PAPERS READ BEFORE THE CHEMICAL SOCIETY.

1.—Modifications of Double Sulphates.

By Spencer Umrrevitte Pickertnec, M.A. Oxon, Professor of
Chemistry at Bedford College.

In a previous communication to this Society (Chem. Soc. J., Trans.,
1884, 686), attention was drawn to the great discrepancies which
exist in the determinations of the heat of dissolution of anhydrous
potassium magnesium sulphate, it being, according to Thomsen,
10,602 cal., and according to Graham only 7000 cal. An investiga-
tion of the corresponding copper salt, CuK,(SO,)2, led to an explana-
tion of these discrepancies.

Potassium Copper Sulphate.

This salt may be prepared without any difficulty by mixing hot
concentrated solutions of the constituent sulphates in equivalent pro-
portions, and allowing the mixture to cool. The liquid should not be
heated to within 20° or 30° of its boiling point, otherwise an abundant
crystalline precipitate will be formed, which, according to Brunner
(Pogg. Ann., 15, 476), consists of a basic double salt having the
composition CuK,(SO,)2,2CuSO,,Cu0,4H,0, decomposable by water,
leaving an insoluble residue of the tetrabasic copper sulphate,

CuS0O,3Cu0.

A considerable quantity of the crystallised salt, CuK,(SO,)2,6H.0,
having been prepared in the above manner, the crystals were
powdered, washed, and dried by exposure to air, after which they

VOL. XLIX. B

2 PICKERING : MODIFICATIONS OF DOUBLE SULPHATES.

were found to contain the theoretical percentage of water within
experimental error.

When heated at 100°, this hydrated salt parted quickly with the
whole of its water, leaving the anhydrous salt in the form of a blue
powder as dark in colour as the hydrated substance itself: when,
however, this blue salt was heated to a temperature of 150—200°, it
was found to lose its colour and become white, or very nearly white ;
this white modification in its turn underwent a change when the
temperature was further raised to about 300° or 400°, and became
again blue or bluish-green. This third modification remained
apparently unchanged by any additional increment of temperature
till the melting point of the salt was reached at a low red heat. The
molten sult forms an opaque green liquid, whick solidifies to form a
glassy mass; as soon, however, as the temperature falls sufficiently
low, this glass suddenly crumbles into an opaque blue powder
resembling in every respect the blue modification obtained at 400°.
The behaviour of the fused salt has been noticed by previous
observers (Thomsen, J. pr. Chem., 18, 35), but the white modifica-
tion has hitherto escaped observation; indeed, when operating on
small quantities of the salt, it appears difficult to obtain it
at all.

It is necessary to cool the fused or strongly heated salt as quickly
as possible, for, if cooled slowly, it will often revert partially to the
white form, and it is, moreover, necessary to fuse it in quantities not
exceeding 5 or 6 grams, in order to avoid any decomposition, as it
begins to evolve sulphuric anhydride at a temperature very little
above its fusing point.

The curious changes of colour experienced by this salt suggested
its existence in distinct modifications, which might be more fully
investigated by measuring their heats of dissolution.

Various specimens were, therefore, prepared and examined in this
way. The results thus obtained are embodied in Table I (p. 6).

The first column in this table gives the number by which the
samples were designated for the convenience of future reference.
Nos. 3415, 3424, and 3425 were obtained from a totally different
preparation of the hydrated salt to that from which the others were
obtained.

w = weight of salt taken.

W = the water equivalent of the calorimeter and its contents, the
volume of the water in it being 601,678 + 0°016 T° c.c., which,
according to Berthelot’s simplified method of calculation, is reckoned
as being equivalent to the same number of grams with a specific heat
of unity (see Mécanique Chimique, 1, 190).

Column VII records which thermometer was employed in the

PICKERING: MODIFICATIONS OF DOUBLE SULPHATES. 3

experiment ; the temperature 7, ¢, and ¢’ being given in the arbitrary
degrees of these instruments.

7 = temperature of the salt at the moment of its introduction into
the calorimeter. ¢ = the initial, and ¢’ the final temperature of the
calorimetric liquid, corrected for (1) the temperature of the salt

according to the equation il where ¢ is the specific heat of the

W+w’
salt ; (2) the exposure of the mercurial column to the temperature of
the air; (3) the calibration correction.

The molecular heat of dissolution (Column XII) is given by the
c= jew where M is the molecular weight of the

w
salt (333°02), and a the mean value of the arbitrary degrees of the
thermometers in degrees centigrade, this being in the case of ther-
mometer 81, 0°38663°, and with 83, 0°38046°.

The next column gives the initial temperature in degrees centi-
grade.

The figures in the other columns will be explained shortly.

The proportion of water to salt taken was about 800 : 1 mol.

For experimental error, &c., see Trans., 1884, 686.

Some small correction should be applied to some of the numbers
given in this table, owing to the presence of a little basic and insolu-
ble salt which the white specimens always contained. The greatest
quantity which was ever found amounted to 0°2 per cent. (in the
case of the white modification), and would necessitate a correction of
about + 12 cal.; some of the specimens of the third modification
contained a similar impurity also, not exceeding, however, 0°06 per
cent., corresponding to 6 cal.; these amounts are so small in com-
parison with the experimental errors that it was not thought neces-
sary to correct for them. The first modifications as well as the fused
Specimens contained no such impurity.

The majority of the experiments quoted in this table were per-
formed at an initial temperature of 18°25° C., and, confining our
attention for the moment to these experiments only, it will be seen
that all the blue specimens which were obtained at low temperatures
dissolved with practically the same evolution of heat, namely,
9709 cal.; the white specimens evolve a very much smaller amount
of heat on dissolving, about 62U0 cal. only, though the various
numbers here are not so closely concordant as they are with the blue
specimens, from a cause which will be mentioned shortly ; lastly, the
blue specimens obtained at higher temperatures evolve an amount of
heat differing from either of the others, namely, 8407 cal. There

can be no doubt, therefore, that the successive changes of colour indi-
B 2

equation M =

4 PICKERING: MODIFICATIONS OF DOUBLE SULPHATES.

cate the formation of distinct modifications of the salt, which may be
conveniently designated as a, B, +.

The method of preparation of the different specimens was varied as
much as possible. No difference is made in the nature of the 2-modi-
fication, whether it is obtained at the lowest temperature at which
dehydration is complete, or at the highest temperature (about 130° as
indicated by a thermometer in an air-bath*) which can be employed
without risking the formation of some of the white salt. The various
white specimens were obtained either direct from the hydrated salt, or
from the blue anhydrous modification, and at temperatures ranging
from 180° to 220°, but in all cases they yielded identical numbers,
showing that this modification also has a perfectly definite existence.
The second blue or y-modification, which begins to appear at 250—
300°, is in a similar manner perfectly stable thronghout a considerable
range of temperature, and even after fusion yields the same numbers
on dissolution. The specimen, No. 2737, which alone gave numbers
lower than the others, was known to contain some of the white modi-
cati n unaltered.

The heat of dissolution decreases, of course, and decreases rapidly,
with the temperature of the water, and we can examine those results
obtained with the B- and y-modifications at temperatures other than
18°25°, only by comparing them with the heat of dissolution of the
a-modification at identical temperatures. Where M, and Mg repre-
sent the heats of dissolution of the a- and B-modifications respectively
at T°, the heat of formation of the B- from the a-modification at
that temperature is M,—Mg, The values of M, at the necessary tem-
peratures are given in Column XIV, having been deduced from
numerous experiments, the details of which it is not necessary to give
here, and the values of (M.—Mg) and (M,—M,) are given in
Columa XV. These values, the heat of transformation of the a- into
the B- and y-modifications will, however, not be constant quantities,
unless the specific heats of all the modifications are identical. Where
this is not the ease, the heat evolved, Q’, in any chemical reaction at
T’, may be calculated from that evolved, Q, at any other temperature
T by means of the equation Q—c’(T—T’) = Q'—c(T—T’), in which c
represents the sum of the specific heats of the reacting substances, and
e’ that of the substances formed (see Berthelot, Méc. Chim., 1,
105). It was necessary, therefore, to determine the specific heats of
the salts in question ; the details of these determinations will be more
conveniently given elsewhere, the general results only being here

* In such a case, the thermometer being placed above a dish containing some
100 grams of the salt, the lower portions of the salt get heated, no doubt, 20° or
30° above the thermometric reading.

VS |S Rm Dw ss G

PICKERING: MODIFICATIONS OF DOUBLE SULPHATES. 5

stated. The molecular heats of the three modifications were found
to be—
= 56°(25
58°735

Substituting these values for ¢ and c’ in the above equation, the
results obtained at the various temperatures were all reduced to 18°25’,
and entered in Column XVI of the table.

In the first place with regard to the y-modification, the numbers
thus obtained are not identical, as they should be, but show a regular
increase with the temperature; it is difficult to account for this fact
otherwise than by assuming that the specific heat of this modification
is considerably greater at these temperatures (8—23°) than experi-
ment showed it to be at 8—43°. In order to render the figures in the
last column identical, the value of cy would have to be 70 instead of
58°7, or at any rate the difference between the two specific heats c, and
cy would have to be five or six times greater than the determinations
gave it. Such a supposition is scarcely admissible, and it will be pre-
ferable to take those experiments only which were performed at
18:25°, and which give + 1302 cal. as the heat value of the trans-
formation of the «- into the y-modification at that temperature.

The numbers in the last column, which refer to the formation of
the white modification, show also a considerable variation, but here
the variation may be accounted for without difficulty.

Berthelot has shown (Ann. Chim. Phys. [5], 29, 295, et seq.) that
most double as well as single salts after fusion do not at once attain
their normal and stable condition as regards their thermal properties,
and it seems not improbable that a similar period of instability may
intervene after they have been heated, even without fusion (s-e
Trans., 1885, 99), especially if, as in the present case, this heating has
been accompanied by soine distinct molecular rearrangement. Unfor-
tunately, at the time when these experiments were made, now nearly
two years ago, Berthelot’s results had not been published, and the
intervals which had elapsed between the preparation and dissolution
of a sample were not accurately noted; the dates which are given in
Column IV are, therefore, for the most part approximations only;
but, nevertheless, they show satisfactorily that the white samples
dissolve with an evolution of about 230 cal. more after they have
been kept for some days than they do when freshly prepared.

Taking the different samples separately—

No. 2739 dissolved after three days with a heat evolution which
gives the value of Ma—Mg to be 3590 cal. (Expt. 9); after
10 days the value (calculated for 18°25°) was reduced to

PICKERING: MODIFICATIONS OF DOUBLE SULPHATES,

3326 (Expts. 16 and 17) and 3296 (Expts. 7 and 8); after
130 days, to 3396 (Expt. 14) and 3393 (Expt. 15): an average

26 2 9 -
reduction of 4 264 + 294 aa 194 + 197 _ =) 23% cal.

. 2750 after three days gave the value 3476 (Expt. 12); after
10 days, 3299 (Expts. 18 and 19); and after 130 days, 3284
177 ; 192 =)

(Expt. 13), showing an average reduction of (

185 eal. in this value.

Yo. 3425 dissolved after one day gave 3602 (Expt. 20); and after
60 days, only 3271 (Expts. 21 and 22): a reduction of 331 cal,
2716 dissolved after one day gave 3528, i.e., 200 cal. above the
average of other samples dissolved after keeping.

. 3415 alone showed no change on being kept 44 days; but as it
gave only 3360 cal. as the value of the reaction when dis-
solved immediately after preparation, it would appear as if
some. peculiarity in the details of its treatment had reduced it
at once to that state in which this modification remains stable.

Summing up these results, it seems evident that some change takes
vlace which is complete in about 10 days, and which results in the
salt dissolving with an evolution of about 230 cal. more than it does
when freshly prepared, the actual numbers being 6159 cal. at first,
and 6489 cal. eventually, the transformation of the a- into the
B-modification evolving 3550 and 3220 cal., according as either the
former or the latter of these above numbers is taken; preference
should be, perhaps, given to the latter, as it corresponds to the more
stable condition. This number will therefore be adopted.

A comparison of Expt. 6 with Expts. 1 to 5 shows that tie a-modi-
fication undergoes no change of this description; and a comparison
of Expt. 27 with 28, and of 34 with 31 to 33, shows the same fact with
regard to the y-modification. Expt. 15 proves that prolonged heating
of a white specimen, at temperatures below that of its preparation,
does not induce any change in it.

The heat development on the passage of one modification into
another, as given above, is calculated, of course, on the assumption
that all the three modifications, when dissolved in water, form iden-
tical solutions, an assumption which the appearance and behaviour of
the solutions, as well as the absence of further thermic effects, fully
justifies.

The heat of formation of the double salt from its constituent sul-
phates, 2, will be given by the following equation :—

N=2+ Mme — (Mx + Meu),

Taste I.—Heat of Dissolution of Anhydrous Potassium Copper Sulphate, CuK,(SO,), = 333°02 ; = mol. = 12

Il.

Colour.

Ill.

Method of preparation.

IV.

Date of
experiment after
that of
preparation.

VI.

VII.

Therm.

VIII.

744 and
2746 |

oe
8. >2739 ..
9,

2743, 2751,
and 2752

eee

Nearly /

white

Quite white {

Very nearly
white

Nearly white

Quite white..

Very nearly
white

i omaines

Crystals dehydrated at 100°; sifted
and reheated at 100—130°.

Crystals heated at 50°; sifted and
reheated at 100—120°.

Crystals heated at 80°, and then
at 115°.

} About lday .. {

} » Sdays.. {

99 Blogg eee

» 120 days..

614°12
614 °12
614°12
614°21
614°21

614 °21

Crystals heated at 215°; sifted and
reheated at 195°.

after being sifted.
| Crystals heated at 180°; sifted and
| reheated at 195°.

| Some of 2739 after being reheated
at 100—120° for 19 hours.

Crystals heated at 180° ..

Crystals heated at 200° ..

A blue specimen heated at all q

} (bout 10 days {
1 day .. {

rp?
\ 3days ..
; 130 days..
130 _,,
130

1 day oc cccccccece
60 days .

IE vcctinn

44 days .

614 ‘27
614°27
614 °12
614 °12
614°12
614 °21
614°21
614°21
614°21

614°03
614 °12
614°03
614 °12
614°06
614°08
614°08
614° 02
614°02
614 02
614°02

‘a ae ee

”

| A white specimen heated over a
burner ; not homogeneous.
Heated below redness ....00++++4
Nearly fused.....scececeees cove
Partially fused...esecesecesesess
Fused......

Pee ss vs (CAC Rdeeetaenee sere

re ee ee ee

About 10 days ....

D ROU incocececs
} About 1 day .. {

614°:

614°
614°
614°
614°
614°
614°
318 °37
614 °03
614°03

TX.

21 °456
21°3845
21 °397
21°7795
21°705

22 -3735

33 ‘781
33 °8205
21°552
21°452
21 °3425
21°651
22 °4685
22 °352
22 °093

32 °598
32 *494
32°473
32 °550
23 *202
23 *2055
23° 2065
17 °3255
17°595
17 *442
17 *432

34°9025
2 5235

22 °303
22 627
23 °483
23 °3575
3 °087

3 °542
33 °409
3 °429
3°494
3984

872

“443

"945
"032
‘066
354
"194
‘190

34°230

33 °861
21 °252
21°7165
21°210
21°709
21°538
22°310
18 *203
18°213

3 °6395

5 +246
3°523
3°004

‘556
"046
"844
"545
*265
*282

‘091
‘O82
‘9615
‘991
‘9605
‘976
*0145
‘0055
‘994,

0°944.
0°915
0°956
0'944
0*782
0°7395
0° 8255

‘7405

Molecular he¢
dissolutio

9753 | on.=)
9605

9577

6202 ]
6161 f

}

IV.

Date of
experiment after
that of
preparation.

eitted } About 1 day .. {

d and } » Sdays.. {

d then —

120 days..

fo ene I; \bout 10 days {
9° 3 ”
> « & day .. {

J
and | » 3 days
130 days..

About 10 days ....

eS . erererren
Aver (
== } About l day .. {

ecces » ts
, » 3 days

ee ” 3 »”
° 99 GD wn cece
fn Gm of

12 392
12° 434
12°588
12 °605
12 °423

12 °471

bo BO bo DO WO bo DO DS
. ao A

ell dll

DO = bo DS bo bo bo
ore
~]

12 °479

12 °380
12 °608
12 °334
12 626
12 549
12 383
5 *995
12 °536
12 *743

VI. VII. VIII.
W. Therm. T.
614°12 81 18 62
614°12 81 18°52
614°12 81 20°35
614°21 81 26 *d4
614°21 81 25°94
614 ‘21 81 21°65
614. ‘27 81 35°62
614°27 81 35°32
614°12 81 18°45
614 °12 81 17°32
614°12 81 18°20
614°21 81 25 °43
614°21 81 21°57
614°21 81 21°41
614°21 81 21°57
614°03 83 87°37
614°12 83 35 °87
614°03 83 37°92
614°12 83 36°47
614:°06 83 22 *84
614°08 83 22°63
614°08 -83 23°02
61402 83 17 °37
614°02 83 17 ‘28
614 02 83 17°34
614°02 83 17°56
614°27 81. 34°76
614 °27 81 29°19
614°12 81 22°70
614°12 81 22 °82
614°12 81 15 *25
614 °21 81 23°37
614°21 81 23°47
318 °37 81 22 -67
614 °03 83 17°15
614°03 83 17 84

Heat of Dissolution of Anhydrous Potassium Copper Sulphate, CuK.(SO,). = 33:

IX. =
t. t’.
21 456 22-988
21°3845 | 22-8905
21-397 22 ‘948
21°7795 | 23 °3385
21°705 23-222
22-3735 | 23-910
33°781 | 34-872
33 °8205 34°9025
21-552 | 22 +5235
21°452 | 22-413
21°3425 | 22-303
21 °651 22 -627
22°4685 | 23°483
22 +352 23 °3575
22 +093 23 ‘087
32 598 33-542
32 °494 33 409
32°473 33 -429
32 °550 33 °494
23 +202 23 -984
23:°2055 | 23-945
23°2065 24 °032
17°3255 | 18°066
17°595 18 °354
17 °442 18-194
17 *432 18:190
34°230 | 35-6395
|
33°861 | 35°246
21°252 | 23°523
21-7165 | 23°004
21°210 | 22°556
21°709 | 23°046
21°538 | 22°844
22 °310 | 23-545
18-203 | 19°265
18°213 | 19 *282

. 80
3 : ' ‘ace Value of the
+ a Fag XII. XIII XIV e change of the a-
Heat of into the B- or y-
f ft M; T° CO dissolution of the | Modification at ;
; : Molecular heat of : a-modification XV XVI
dissolution. at T° C. T° C 18 25° C
- | :
22-988 1°532 9753 ~
22 -8905 1°506 9577 9665 |
22 948 1°551 9743 | oreo |
23 3385 | 1°559 9781 f°“"" +9709 | 18°25 — — -
23-222 1°517 965
23°910 | 1°5365 9744 J |
| ; 1-09 |
34 *872 ‘091 6847 29Q9 > of 2921@ 296
34°9025 | 1°082 e716 f — ante mene — —_
| 22-5235 | 0-9615 | 6115 |) ( 3590
| 22-413 0-991 | 6202 . Boy
| 22-303 09605 | 6161 ae | | | _—
22 °627 0-976 | 6233 | $18°25 9709 4 — 3476
23-483 1°0145 | 6425 | | - 3284
23°3575 | 1°0055 | 6313 | _ 3396
23 -087 0°994 | 6316 |) UY - 3393
|
33 542 0'944 | 5984 sien | ane —
3: 3326
33 °409 0°915 5680 5807 an sous .
33 -429 0956 | 5898 5834 | as _ 3276 3299
33°494 | 0°944 | 5770 | ” ¥
23-984 | 0°782 | 4848 | 3562 | 3602
23 -945 0-7395 | 5171 ¥ } 9°85 8410 { -
24-032 0°8255 | 65187 6179 | sasl ave
18°066 0°7405 4598 aon
J 3360
18 "854, rf) 7 59 | 4724. 4661 — —_ 3309 y
18-194 | 0-752 | 4697 mane — oe
18-190 0-758 | 4668 _ | 32 3338
| |
bs ms ‘ | a ia —_ . ——
| 35°6395 | 1°4095 8892 |
| 8866 22°98 | 10105 1239 | 1252
| 35 °246 1°385 8840
| 23°523 1271 7972 ns
| 23-004 | 1°2875 | sa74y 8128 |
| 22-556 1°346 8431)
| 23-046 1-337 8426 | . _ ones rv. 1302
| 22°844 1 306 s341 (5407 18°25 o708
| 23 *545 1°235 8430
| 19°265 1-062 6591 —_ 09% ‘ 1472 1444
| 19-282 | 1-069 6526 6558 7935 soso -

= 333°02 ; =

mol. = 12°488 grams.

PICKERING: MODIFICATIONS OF DOUBLE SULPHATES. 7

in which Mxy,, Mg, and Mj, are the molecular heats of dissolu-
tion of the double salt of potassium sulphate, and of cepper sul-
phate respectively,* and N the heat disturbance on mixing solutions
of the single sulphates. This latter quantity was found to be nil,
in accordance with the observations of other physicists, and the
values of My and M,, were found by means of numerous experiments
to be — 6495 and 15633 respectively at 18°25°; the algebraic sum of
these is + 9138, and hence the heat of formation of the a double salt
will be (9138 — 9709 =) — 571 cal. 7.e., it is an endothermic com-
pound. The other two modifications on the contrary would be formed
hy exothermic reactions :—

[K,S0, + CuSO, = B-CuK,(SO,).] = 2649 cal.,
and [K,SO, + CuSO, = 7-CuK,(SO,)2.] = 731 ,,

That all these three salts are in reality compounds, and not mere
mixtures of the uncombined sulphates of copper and potassium, is
clearly shown by the fact that none of them dissolves with the heat
evolution with which the mixed salts would, namely 9138 cal., and
that those two modifications which approach most nearly to this quan-
tity are both blue substances, whereas a mixture would be white;
the only modification which is white dissolves with an evolution of as
much as 2649 cal. less than a simple mixture would do.

It remains only to be added that whereas the first change under-
gone by this salt, the passage of the a- into the B-modification, is an
exothermic action evolving 3220 cal., the second change, that of the
B- into the y-modification, is an endothermic action, absorbing 1918
cal., as measured at 18°25°.

The mean of two determinations of the heat of dissolution of this
salt which were made by Thomsen gave 9396 cal. at 17°27°, whereas
at this temperature the «-modification, according to my experiments,
would evolve 9617 cal. on dissolving, indicating that Thomsen’s speci-
men had been overheated, and contained some of the S-modification.
The experiments, however, are not strictly comparable with mine,
as he used a proportion of water amounting to only two-thirds of
that used by myself.

Potassium Magnesium Sulphate.

An investigation of potassium magnesium sulphate conducted on
the same lines as that of its copper analogue was found to be com-
plicated by various circumstances.

* If the single sulphates are dissolved in 400 molecules of water, the double

sulphate must be dissolved in 800 molecules in order to make this he renee correct.
These proportions were taken.

PICKERING : MODIFICATIONS OF DOUBLE SULPHATES,

In the first place, considerable difficulty was met with in preparing
the salt itself in a state of purity. The double salt is considerably
more soluble than its component sulphates, and, therefore, unlike the
copper salt, it does not crystallise out on mixing saturated solutions
of the two sulphates. The mixed liquids should be evaporated
while hot, and then allowed to cool ; spontaneous evaporation should
not be resorted to, and the magnesium sulphate should be present in
considerable excess, otherwise the crystals separating out will con-
tain a large quantity, or even consist entirely of uncombined potas-
sium sulphate. Moreover, since the double salt is entirely decom-
posed by an excess of water, the crystals should be washed free from
mother-liquor only by means of a saturated solution of the double
salt itself. After many unsuccessful attempts, a large quantity of
the crystallised salt was prepared containing the theoretical per-
centage of water; from this, the anhydrous specimens were obtained.
But here another difficulty arose. Just like magnesium sulphate
itself, the double salt does not lose its water at the same low tempera-
ture, and with the same ease as the copper compound does. The
whole of its water is not evolved below 155°, and, as it was feared
that this temperature would be sufficiently high to induce some
change in the constitution of the salt, the same expedient was
adopted as in the case of magnesium sulphate (Trans., 1885, 101),
namely, dehydrating the salt as far as possible at the required tempe-
rature, and making a correction for any residual water which is still
retained. This correction was made in the same way as for magne-
sium sulphate, necessitating a knowledge of the heat of dissolution of
the hydrated salt as well as that of the anhydrous salt at some par-
ticular temperature. The former of these quantities was found to be
— 9777, and the latter, as will be shown below, + 12037 at 22°28°,
and, consequently, the correction to be applied to the numbers obtained
with a sample retaining, for instance, 0°6 per cent. of the total water
present when fully hydrated would be re of (12087 + 9777) cal. The
various corrections are given in Column X of Table II (p. 9), while
the method of preparation and other details are given in the last
column. All the experiments in this case were made at the same
temperature, 22°28° C., with the exception of those with sample
No. 3432, where the results at a higher and a lower temperature gave
the means of calculating what the number would have been obtained
at 22°28°. The symbols used are identical with those in Table I. The
sp. ht. of the salt was taken as 0°17.

To begin with, we are met by the same fact observed in the case of
the copper salt ; that all the specimens prepared at low temperatures,

ranging in this case from 110° to 155°, are identical in nature, and give

“spiem
*posng
*“spiem
-194jB SINOY Z PIajossiqy “posng
*SplBM
*pasny
*pasny
*pasny A]perseg
“q8voq pol 8B eV ”

10338 rea [ poalosstq

-1aye skep F paalossiq,

“oOLT 98 posedoig

Aun

‘snoipfyue {cc 3e powedaig

*1978M *4U99 Jad 62.0
{,081 38 poiedoarg

“108M “U0 19d 09.0
'.081 98 paiedaig

poureyor

pouyezo1

*“19yOM
18303 943 Jo “qua0 aod 76.
pourejol {°O (OIL 38 powedaig

£208
Iehl
SIL
9RLL
6F6L
OFt8
1026

98601
aIlIt

> LEOZT

OTOIT
CPrsol

82-23

qe Zura3

Prtbddted
Prd deed

-1Z 38 81801
FS Ye CLELT

oo
el
1S

©.0
a
A

Gl1t8-1

§ 6F6IT

oo
26081 2 so1Z1

eSOZl

0
0
18 +
eo +
OZI +

Llb +

*syVulsyy
“IX

*paqoo1109 JT

‘1X

*U0T}DILIOD
=

TL0z1
Sc6IL

S26IT

§ SF9IT

BOOT 2 Egctt

LSI-T

O8I-T

¢ecoz- IT
Cegl-T
£02- I
622- I
992-T
Goce. T
89F-T
Chl. I
9IL-T
1e@l-T

¢188-T
cle6- 1

6+ T
688- I

O16. T

1¢8. I
Gces- I

F~s
“TITIA

109- 2&

8EF- €E
CILb- SE

608- €&
£OF- £8

9II- $8
£91- 8

PEL. HE
S196. &¢

FES. HE

GREO- FE
G6F6- £8

89U- 68

Scrl-
002-
€aI-
912.
122.
Z91-
L60-
B22-
6¢8-
SO9T-
C822
GIs.

0&2.
$290: 28

P28. ZE

GL8I- Ze
LII- Zé

28- FID

1@- F19

LZ 19
12-19
L2-F19
L@-F19
LZ $19
LZ-F19
L2-F19
12-19
1Z-F19
16-19
0&- F19

9F6- OI

F10- TI

110- IT
1g6- OL
OLI- IT
060- IT
StI- IT
ILI-1T
8eI- I
010- IT
Le0- IT
6eI- I
0F2- IT

18
18

12. F19
LZ. B19

12- F19
12-419

1Z-F19

12-F19

120-11
OFI- IT

T10- If
6F0- IT

P8I- IT

¥60- IT

42-¥19 | 290-11

#008
+908

%
‘TA

| *ULLOUL,

‘Al

‘ajdueg
‘I

‘SUIBIS ET).T[T = ‘jou ry
"£9-£66 = *COS) MSN ‘eppydjng wnseubopy unissnjog snouphyup fo uornpossig fo yoayT—[[ A14VJ,

ah
oo
ee
<
se)
=
4
=)
mM
=
4
a
=)
jo)
=
=
oO
mM
A
2)
—
i
<<
oO
—_
=
Lm
a
2)
a
o
“A
(=
-
oO
=
=

10 PICKERING: MODIFICATIONS OF DOUBLE SULPHATES, |

the same evolution of heat on being dissolved, namely 12037 cal. at
22°28°, the numbers obtained being very concordant, when it is remem-
bered that a correction depending on the accurate determination of
a very small quantity of water had to be applied to most of them.
155° appears, however, to be very near the limit within which such
results are obtained; when heated to 170° the samples show a con-
siderable decrease in their heats of dissolution, just as in the case of
the copper salt, and this decrease becomes rapidly greater as the
temperature employed is higher, till it reaches a climax in the fused
specimens, of which the heat of dissolution is only about 7400 cal., or
not much more than one half of that of the specimens prepared below
155°, and which may be termed the a-modification.

No modification intermediate between the a-modification and the
fused salt appears here, as is the case with the copper salt, but it would,
I think, be rash to say that such does not exist. It is a character-
istic of the magnesium salts, in contradistinction to the copper salts,
that their dehydration, and the various changes which they undergo,
not only take place at higher temperatures, but are merged one into
another, so that it is often difficult to obtain one of the products free
from a certain amount of the others (see 'T'rans., 1885, 101). Such is
very probably the case with the double salt. It may be either that
an intermediate modification is formed between 155° and a red heat,
but that the range of temperature throughout which it is obtainable
is so small that the conditions of the present experiments never
sufficed to produce it unmixed with the other two modifications (in
which case the fused salt should be termed the third or y-modification),
or else it may be that the temperature at which the changes take
place being so much higher than they are in the case of the copper
salt, the very highest temperature (a bright red heat) which the salt
is capable of bearing without decomposition is high enough to form
only the second instead of the third modification ; if this be so, the
fused salt should be termed the f-modification. The latter! view is,
perhaps, preferable, seeing that the heat of dissolution of this last
modification is so much smaller than that of the first one, bearing
towards it about the same ratio as the B- does to the a-potassium
copper sulphate.

The fused salt exhibits two peculiarities which the unfused samples
do not. When thrown into water, it dissolves with extreme slowness,
it does not cake, but forms a milky liquid which becomes clear only
after the lapse of about 25 or 30 minutes. This has also been noticed
by Berthelot (Ann. Chim. Phys. [5], 29, 329), and resembles, though
in an intensified form, the behaviour of monohydrated magnesium
sulphate. The second peculiarity is that when first put into water, a
fall of nearly 01° precedes the rise in temperature: this, no doubt,

—— ah «at of. SR te

PICKERING: MODIFICATIONS OF DOUBLE SULPHATES. 11

is due to its passage into the a-modification previous to its disso-
lution.

The length of time required for the dissolution of the fused speci-
mens renders the heat determinations very difficult ; indeed, these can
be regarded as approximations only. Omitting Expt. 18, which
appears to be exceptionally high, the mean of the last five experiments
gives 7431 cal. as the heat of dissolution of B-MgK.(SO,)2, and the
value of the transformation of the a- into the #-modification as
+ 4606 cal., measured at 22°28°.

According to my own experiments, potassium sulphate dissolves in
400 mols. of water, at this temperature, with an absorption of 6200 cal.,
and magnesium sulphate, under like conditions, with an evolution of
20722 cal. This gives the following values for the heat of formation
of the double salt from its component sulphates :—

[MgSO, + K.SO, = a-MgK,(SO,).] = 2485,
(MgSO, + K,SO, = B-MgK.(SO,).] = 7091,

both reactions being exothermic.
Berthelot (loc. cit.) determined the heat of dissolution of a specimen
of the fused salt; his results when reduced to 22°28° become—

(1) Dissolved at once . 8039 cal.

(2) sith atieaill 7132 ,
(3) s after 3 weeks .... 6432 %

(4) - and finely powdered 5588 ,,*

from which he concludes that after being kept, this salt dissolves
with a decreased evolution of heat, due te its parting with some of
the heat it had absorbed during fusion, and further, that powdering
facilitates this loss. _My own experiments, on the contrary, if they
show any change at all in the behaviour of the salt when kept, show an
increase in the heat of dissolution (comp. Expt. 18 with 17, and 20 with
19), a change similar in nature to that which certainly takes place with
the B-copper salt.t The difficulties of experimenting in this case,
however, are so great that much weight cannot be attached to these

* Berthelot’s actual numbers are—(1) at 17°, 7300 cal. ; (2) and (8) at 20°1°, 6880
and 6180 cal. ; (4) at 20°8°, 5421 cal. Taking with these my own determinations of the
heat of dissolution of potassium sulphate and of magnesium sulphate at these three
temperatures—which are with the former — 6600, —6340, and —6300, and with the
latter 20383, 20610, and 20654 cal. respectively—we get for the value of the equa-
tion [MgSO, + K.SO, = B-MgK,(SO,)2] 6483, 7390, 6432, and 5588 cal. in the
four experiments, and the difference between these quantities and the sum of the
heats of dissolution of magnesium and of potassium sulphate at 22°28° produces
the numbers given in the text.

+t Neither the fused copper nor any specimens of the y-modification showed any
change on keeping.

12 PICKERING : MODIFICATIONS OF DOUBLE SULPHATES..

differences, and it will be sufficient to remark that, on the whole,
Berthelot’s experiments are fairly concordant with my own, the mean
of his giving 6798, or that of his first three 7201] cal.* against the
mean which I have taken for mine, 7431 cal.

The salt examined by Thomsen (J. pr. Chem., 18, 27), dissolved
with an evolution of 900 cal. less than that of the «-moditication,
showing that this preparation was either not anhydrous or that it had
been over-heated and partially converted into the 8-modification.

The great insolubility of anhydrous alum points, I think, to the
probability of its being in reality different in constitution from the
hydrated salt; unfortunately this very insolubility and the difficulty
with which it parts with its water renders any experiments similar to
the present ones impossible. It appears to be a moot point whether
anhydrous alum is absolutely insoluble or not; so far as my experi-
ence goes it is certainly not, but at the same time I have found it
quite impossible to deprive alum entirely of its water without causing
it to lose a very appreciable amount of sulphur trioxide. Pure
anhydrous alum is probably unknown.

II.—Modifications of Double Sulphates.f Part II. Specific Heat
Determinations.

By Srencer Umrrevitie Pickerine, M.A. Oxon., Professor of
Chemistry at Bedford College.

Douste sulphates of the type M’M’'(SO,).,6H,0 were originally
regarded by Graham (Phil. Mag., 6, 327, 417; 10, 216, &.) as
derived from the corresponding heptahydrated magnesian sulphates
by the displacement of one of the molecules of water contained in
them by one molecule of an alkali-metal sulphate. Thomsen (J. pr.
Chem., 18, 29) controverted this idea by attempting to prove that the

* All the specimens used by me were finely powdered, but the powdering was
performed as soon as the fused salt had solidified and while it was still quite hot. It
is possible that a slight absorption of moisture during the powdering of Berthelot’s
specimen, while cold, may be the cause of low results of Expt. 4. The proportion
of water used by Berthelot was smaller than that used in my own experiments, and
hence an absolute concordance of results cannot be expected.

+ This communication formed part of a paper entitled ‘“‘ Notes on the Constitu-
tion of Hydrated and Double Salts,” which was read before the Society. The
theoretical portion of this paper will be published hereafter.

nan fF fF, se ss, == ~~

PICKERING : MODIFICATIONS OF DOUBLE SULPHATES. 13

heat of combination of the various water molecules in the double sul-
phate was very different from that of the water molecules in the cor-
responding single sulphate. I have already shown, however (Trans.,
1885, 102), that the experimental data on which Thomsen relied were
incorrect, and I shall elsewhere endeavour to prove that his argument
was founded on a theory which is quite inadmissible.

Substitution or displacement is one of the many ideas in chemistry
which do not admit of any exact definition, and we can only settle
whether one substance is a true substitution-product of another by
general considerations as to the mode of its furmation, and the extent
to which it retains the fundamental characteristics of the parent sub-
stance.

The case now under consideration is, I think, as true an instance of
substitution as any which can be adduced. The manner in which the
double sulphates are obtained is of the simplest character, and they
exhibit the peculiarities of the sulphates from which they are ob-
tained, even to the most minute details.

The heptahydrated sulphates lose 6H,O at about 100°, leaving the
monohydrate M"SQ,,H,O, whilst at the same temperature the double
salts part with their 6H,O, leaving the salt M”SO,,K,SO,. In the
case of the copper compounds, both of these are of a light blue colour,
and both, when raised to nearly 200°, are decomposed, forming white
substances, the anhydrous salt in the one case, and some differently
constituted double salt in the other. The peculiarities which distin-
guish magnesium sulphate from copper sulphate are accurately repro-
duced in the double salts which it forms; the monohydrate is not
obtainable at 100°, but requires a temperature of 150—160° for its
formation, and begins to suffer decomposition at a few degrees
higher; in like manner the double salt does not part with its water
below 155°, and begins to pass into another modification at 170°; the
monohydrate cannot be completely decomposed at a temperature
below 250—300° instead of 200°, as in the case of the copper salt,
and the double salt also requires a temperature considerably higher
than 200° to change it into the second modification.

Person was the first to show that the specific heat of a hydrated
salt is equal to the sum of the specific heats of the anhydrous salt
and of the water contained in it, reckoned as solid water. Subsequent
experiments have invariably confirmed this view, and, though I have
reason to doubt its absolute correctness, it is probable that no direct
determination of specific heats would be sufficiently accurate to detect
any flaw init. Now, if it were found that the specific heats of the
double sulphates resembled those of the hydrated sulphates in being
equal to the sum of those of their constituents, such a fact would lend
strong support to the view that they were similar in constitution to these

14 PICKERING: MODIFICATIONS OF DOUBLE SULPHATES.

hydrated salts. An investigation of the specific heats of the three
modifications of anhydrous potassium copper sulphate was, therefore,
undertaken.

The method employed in these determinations consisted in dis-
solving a portion of the salt in water under precisely the same con-
ditions as for the determination of the heat of dissolution, except that
the temperature of the salt, instead of being nearly identical with
that of the calorimetric water, was about 35° higher. The salt to be
dissolved was weighed out into a short wide test-tube fitted with a
small thermometer; this tube was fixed in a short-necked flask nearly
full of water, the whole being heated in an air-bath to the required
temperature. The flask with the water acted as a jacket to the tube
containing the salt, thus allowing its removal from the bath to the
calorimeter without any appreciable alteration in temperature during
the few seconds so occupied. This method cannot of course claim any
of the accuracy attained by Pape, Kopp, and others using very elaborate
apparatus, but it was found to yive results sufficiently accurate for
the present purposes.

The specitic heats of salts as given by different authorities differ so
much that it was thought advisable to make determinations of those
of copper and of potassium sulphate, at the same time, and with the
same apparatus as the double salts. The results are given in the
accompanying table (p. 16), where the symbols used are similar to those
in the tables in the previous communication. In each case the same
sample was used both for the determination of the specific heat and
of the molecular heat of dissolution, M.m is the molecular heat of
dissolution in the specific heat experiments without any correction for
the salt temperature 7, which, in these cases, is given in degrees
centigrade, as also is the interval through which the salt was cooled
(r—t). The other temperatures are given in arbitrary degrees as pre-
viously. c¢ is the specific heat, water = 1; and c» the molecular
specific heat.

The three modifications of potassium copper sulphate, it will be
seen, are as clearly distinguished by their different specific heats as
they are by their different heats of dissolution. The a gives 56:025,
or 0°168, the 6 51°24, or 6°154, and the y 58°735, or 0'176. Now the
sum of the specific heats of copper sulphate and potassium sulphate is
54°46, or 0164, but it is probable that this is somewhat smaller than
it should be, for, owing to the powdery nature of anhydrous copper
sulphate, some particles of it often float on the calorimetric water
before they sink, and these must consequently part with some of their
heat to the air, thus giving too small a value for the specific heat of
the salt: were it not for this, the sum of the specific heats of the
mixed sulphates would no doubt approach more nearly to that of the

rr Tr a a a a a ae a

~~ aS lle Oa la

PICKERING : MODIFICATIONS OF DOUBLE SULPHATES. 15

a-modification of the double salt than it does. Even as it is, however,
the numbers are sufficiently concordant* to show that if any of the
modifications has a specific heat nearly corresponding to that of a
mixture, it is the a-, and not the B- or y-modification. It is in the a,
therefore, if in any, that the potassium sulphate present appears to
perform the same function as water in the hydrated single salt. Such
a view is in full accordance with the fact that the a-modification
is obtained from the crystallised salt at the lowest possible tem-
perature, and that this latter in its mode of formation and general
characteristics closely resembles the fully hydrated single sul-
phate.

It will be noticed that the heat of combination of MgSO, + K.SO,
= 2485 cal., whereas MgSO, + H,O = 7016 cal.; and again, that
CuSO, + K,SO, = — 571 cal., whereas CuSO, + H.O = 5143 cal. ;
the molecule of water which is combined with considerable energy
being displaced by a molecule of potassium sulphate which combines
with but feeble energy. It must be remembered, however, that it is
the hydrated, and not the anhydrous salt which is prepared from the
mixed sulphates, and when we take the heat-formation of this we find
that in the case of the copper salt, whereas the sulphates K,SO, and
CuSO,5H,0 in crystallising out separately would evolve 9212 cal., the
double salt would evolve 13,728 cal., and hence the facility with which
it may be formed, and its perfect stability when formed; the magne-
sium salt, however, cannot be formed with the same ease, and is
entirely decomposed by excess of water ;+ this is due to the fact that
in erystallising out it would evolve only 9851 cal., a quantity smaller
than that which MgSO,, 7H,0, and K,SO, would separately, 10,310
cal. Its formation is possible only in the presence of a large excess of
magnesium sulphate, or else at a high temperature, where the relative
values of some of these quantities are no doubt changed.

The absence of any thermal disturbance on mixing solutions of the
component sulphates is generally taken as indicative of the non-
existence of the double salts in solution. In the case of the potassium
magnesium sulphate this may be so, for this salt is entirely decom-
posed by water, but it can scarcely be so with the copper salt, for, if
it were, it would be difficult to see why any of it should separate out
at all on mixing the solutions, unless it be regarded as an insoluble

* Quite as concordant as they are found to be with hydrated salts.

+ A sample of CuSO,,K,S0,,6H,O, which gave on analysis 24°48 per cent. of
water (theory 24448) after being washed continuously till three quarters of it had
been dissolved, contained 24°456 per cent., ¢.e., the same as before the washing. A
sample of the magnesium salt containing 26°809 per cent. of water (theory 26°822),
after being similarly treated was found to contain only 18°89 per cent., showing
that it contained 30 per cent. of uncombined potassium sulphate.

16 PICKERING : MODIFICATIONS OF DOUBLE SULPHATES.

substance, a view which is hardly reconcileable with its composition
and properties.*

By means of the specific heats, the heat of formation of the various
modifications of the copper salt may be calculated at the temperatures
at which they are actually formed.

The conversion of the a- into the A-salt at 18°25° C. = 3220 cal., and
at 200° it will be 4089 cal., as given by the equation—

H — ca(T’ — T) = H' — ca(T’ — T) (see p. 7),
the heat of its formation, and therefore the tendency which it has to
form increases with the temperature.

The transformation of the 8- into the y-modification is an endo-

thermic reaction, absorbing 1918 cal. as measured at 18°25°, and its
being such accounts, no doubt, for the ease with which it reverts to
the f-modification unless cooled rapidly (see p. 2), whereas the
f-modification being formed in an exothermic reaction, exhibits no
tendency to revert to the a-modification from which it is formed (see
p- 7).
But here an anomaly arises: at about 300°, the lowest temperature
at which the y-modification appears, its formation will be even more
endothermic than at 18°25°; applying the equation given above,
mutatis mutandis, it will be found to be — 4031 cal.t when formed
from the 8-modification, or 459 cal. when formed from the a-modifica-
tion, the formation of the B- from the a-modification at this same tem-
perature being 4490 cal. In other words, although the tendency to
form the f-modification (as measured by the heat development) in-
creases continuously with the temperature, and the tendency to form
the y-modification decreases with the temperature, yet a certain rise of
temperature converts the former into the latter. The only conclusion
which can be drawn from this is that the heat development is not the
determining factor of this change. We have here a striking excep-
tion to Berthelot’s “ Principe de Travail Maximum ” (Méc. Chem., 2,
417), which has of late been attacked to no inconsiderable degree.

* Saturated solutions of copper and potassium sulphates were mixed at 12°.
About 48 per cent. of the sulphates present crystallised out in the form of the
double salt. On performing a similar experiment with the magnesium sulphate, no
crystallisation took place till about 45 per cent. of the water had evaporated, and
then it was the wuncombined salts which separated out. Zinc and nickel sulphates
resemble copper sulphate in this respect, whereas those of iron and manganese
resemble magnesium sulphate. The formation of the double salt in the liquid
appears to be a process requiring a considerable time. In one case, it was found
that only 60 per cent. of the total crystals yielded had separated out after 12 hours,

and in another case, after 24 hours, 95 per cent.
+ According to the experiments on the heat of dissolution this number should be
even greater than when based on the direct determination of the specific heats

(see p. 7).

Specific Heat Determinations.

Sample.

t’ —tor
t—t’.

7222 at

10°35° C.

‘59 — 10°35 =) 32°23
2°98 — 10°34 =) 32 64

15687
15676 | 15670 at
15655 { 10°7° C.
15644

15670
15670

(44°75 — 10°68 =)34:07
(43 ‘80 — 10°745 =)33°055

a-OuK(SO,)>.
13. 3426

14. ,,

15. ,,

20°1885

20 °571
20 *324

8010 at 821°C.

8016
80L6

‘23 — 8°24 =) 35°99
‘39 — 8°24 =) 34°15

B-CuK,(SO,)>°
Wins |
17} 3425

18. ,, |
19. ,,

See Expts.

12-640
12-732

21 and 22

614 ‘08
614°08

in Table I.

83
83

23 *2345
23 °225

24°3495
24 °356

5179 at 9°86° C.

5179
5179

‘34 — 9°86 =) 33°48
‘05 — 9°86 =) 83°19

y-CuK,($0,)>. |

20
21 } 3424

22. ”
23. 55

See Expts.

12 320
13 246

35 and 36

614 08
614°08

in Table I.

6558 at 7°935° C.

6565
6565

Ill. An Examination of the Pheno! Corstituents of Blast-furnace Tar,
obtained hy the Alewander,and McCosh Process at the Gartsherrie

Aronworks. «(Part I.)

By Watson Suiru, Lecturer in Chemical Technology in the Victoria
University, Manchester, with Messrs, J. F. H. Courrs and H. E.

BRotrHERS.

A PAPER was read by one of us before the Society of Chemical
Industry (Liverpool Section), December 29th, 1883, “ Ona Preliminary
Examination of Blast-furnace Coal-tar,” from the Gartsherrie furnaces,
in which Scotch coals are used (mostly unsuitable for coke making).
In this paper the following observation was made in the examination cf
the higher boiling oils (see Jour. Soc. Chem. Ind., 2, 495), when the cz
losses by treatment with sulphuric acid and alkali respectively were =a
determined by suitable measurements. The oils of higher boiling point <a
contained more compounds absorbed by acids (basic constituents) and less .;
phenol constituents absorbed by soda, than the oils boiling below them. aw
It was now determined to ascertain approximately what proportions “<<
of crude phenols, and of amido- or basic-constituents, are contained in ©
those portions of the tar-oils analogous to the «arbolic-oil and creosote- mt
oil of ordinary coal-tar. For this purpose, known volumes of the oil &
were repeatedly treated with caustic soda-lye until exhausted 0 |’,
phenol-constituents ; the latter were liberated by acid, collected, an

measured, with the following results :—

Per cent. by
vol. of phenols

\ obtained.
By the Ist treatment with equal vol. of caustic soda of
1:08 sp. gr. ..ccceccccccccccccnces cerccccore 17°5
By the 2nd treatment with 0°65 of their volume of soda
OF 1°15 ap. gr. cc ccccccccccccccacccacsececces 4°6
By the 3rd treatment with 0°75 of their volume of soda
Of 1:20 sp. gr. ...ccccerecccecccecccecececens 0°7

By the 4th treatment with 0°65 of their volume of soda

ee aeeereeer ee eeereere ee eeeeeeeeee

23°1

In order to determine the percentage of basic constituents present,
1160 c.c. of the blast-furnace tar-oils were shaken for a day with
840 c.c. of dilute sulphuric acid of about 1:2 sp. gr. After standing

12 hours, the aqueous layer was separated, and treated, first with a
C

VOU. XLIX,

18 SMITH, COUTTS, AND BROTHERS: PHENOL

certain excess of caustic soda, and then with common salt; after
which it was allowed to stand in a graduated jar. The volume occu-
pied by the layer of basic substances was read off, and it was then
removed by means of ether; the ethereal solution being heated to
expel the ether, and the basic residue again measured. It was observed
that though the odour of the crude bases resembled that of the quino-
line bases, yet it was by no means so marked in these bases as in the
case of the basic oils obtained from ordinary coal-tar oils by a similar
process; 1160 c.c. oils by this treatment gave 150 c.c. of crude bases
= 11:09 per cent. by volume.

These blast-furnace tars are consequently very rich in phenols,
which may be roughly stated to constitute 20 per cent: of their
volume. This proportion far exceeds what can be obtained in a
similar manner from ordinary coal-tar oils (gas-retort coal-tars), and
we would point to this circumstance as being to some extent a con-
firmation of K. E. Schulze’s theory with regard to the probable
formation of, at all events, a considerable proportion of the aromatic
coal-tar hydrocarbons, by the breaking up at higher temperatures of
first-formed phenols into the elements of water and hydrocarbons
(see Annalen, 227, 143). In such case, we should expect to meet
with intermediate tars, which contain, in predominating quantity,
these half-way phenol constituents, if they may so be termed. These
blast-furnace tars (and probably to a less extent the Jameson and
Aitken coke-oven tar-oils) are of this intermediate character.

If Schulze’s theory be correct, amongst these phenols we ought to
be able to find those members most nearly corresponding to the
hydrocarbons which predominate in what we may term normal coal-
tar, produced in gas-retorts at the highest temperatures. For the
purpose of testing this question, we have made a careful examination
of the blast-furnace tar phenols, with the details and results which
follow :—

A rough quantitative fractionation was made of the crude phenols
obtained (Jour. Soc. Chem. Ind., 2, 497) by prolonged agitation
of about 14 gallons of the tar-oils of sp. gr. 0°988 with an equal bulk of
caustic soda-lye of sp. gr. 1°08, and subsequent treatment of the soda
solution with excess of sulphuric acid (2 gallons of dark brown-
coloured phenols of sp. gr. 1°07 being obtained); this showed
that whereas only 5°63 per cent. by volume distilled over between
180° and 210°, no less than 30 per cent. passed over between 210°
and 240°. The results of the fractional distillation were, in fact, as
follows :—

CONSTITUENTS OF BLAST-FURNACE TAR.

Per cent.

Below 190°
BIO—-BO on ceccccccsccs
200—210
210—220
220—230 1384 bs07 per cent.
230—240 10°84
240—250 8°52
250—260 6°80

8:14

270—280 3°65
280—290 2°66
290—300 3°72
300—360 2°88.
360° to coking of residue. 13°50

18:0 per cent.

91°70

The portion which distilled over below 180° consisted chiefly of
water, and was disregarded. The other fractions, which remained
liquid at ordinary temperatures (that distilling from 360° to coking
of the residue solidifies to a red semi-solid resin), were now dried by
remaining in contact with. fused calcium chloride for some time.
Each of these fractions was. then decanted, and submitted to
refractionation, all of such fractions distilling below 300° being
separated carefully into sub-fractions of 5° eaeh, whilst those from
300° to 350° were, as far as possible, separated into fractions of 10°
each. The fractionations were repeated, and the selections made in
the usual way, when it was found that by far the larger portion of the
phenols distilled between 210° and 225°. During the distillation of the
portions boiling below 300°, much sulphuretted hydrogen was at first
evolved, and later on sulphurous acid ; at the same time, a considerable.
deposit of free sulphur was formed in the condensing tubes. This
evolution of sulphuretted hydrogen and sulphurous acid with deposi-
tion of sulphur, is most marked with the lower boiling portions, and’
is evidently due to the decomposition of peculiar sulphur.compounds,
unstable on distillation. After a few distillations, the evolution of
gas ceased, and the phenols then possessed a much pleasanter odour.
In fractionating the portion distilling from 300° to the coking-point
of the residue, it was observed that just at the coking point a peculiar
decomposition occurred, both hydrocyanic acid and ammania being
evolved, distinctly recognisable hy their odours. It is probable that
this may be due to the decomposition of peculiar higher: nitriles in
presence of a limited amount of moisture and hydrogen, ammonia and

c 2

20 SMITH, COUTTS, AND BROTHERS: PHENOL

hydrocyanic acid being formed. It seems strange at first sight that
nitrogenous compounds of this kind should be present in the phenols
after the solution in soda and precipitation with acid ; but even such
a non-basic substance as naphthalene is often found in crude phenols.

For further operations, the final fractions were classified into three
groups: (A) All the fractions boiling below 230°; (B) all boiling
between 230° and 300°; and (C) all boiling above 300°.

Ezamination of the Constituents of Group A.

The fractions in this group were distilled until they boiled pretty
constantly ; each was then placed in a freezing mixture, when that
boiling from 180° to 185° at once crystallised to a solid mass, melting
again at — 3°. The next fraction showed signs of incipient crystal-
lisation at — 13°. The first consisted chiefly of phenol, C,H,;-OH,
but was not quite pure, as the low melting point shows; it contained
small quantities.of meta-cresol. A larger proportion of cresol is pre-
sent in the other fraction refusing to crystallise above —13°. The
examination had not proceeded far before it was found that on
dissolving the various fractions in soda, milky solutions were formed,
indicating the lingering presence of certain impurities. Their removal
was effected by shaking the milky solutions with ether, and separating
the ethereal layer by means of a tap-funnel. Sulphuric acid then
liberated the pure phenols, which were dried over calcium chloride
and refractionated. The-solutions of the higher boiling phenols in
caustic soda had a blue colour, increasing in depth with the rise in
boiling point. So far it was satisfactorily proved that the phenols
boiling between 180° and 200° consisted chiefly of phenol, C.H,-OH,
and the cresols, but that the proportion of cresol to phenol was very
much larger than in ordinary crude phenol similarly extracted from
gas-tar. It may be interesting, however, to mention that the relative
amount of phenol, C,H,-OH, in the crude phenols of blast-furnace tar
considerably exceeds that found in the crude phenols extracted from
the Aitken and Jameson tar-oils (that is, by using for a given volume
of the different tar-oils the same proportion of caustic soda solution of
the same strength). The following table, based on the results of the
fractional distillations of the crude phenols obtained by the same
process from blast-furnace tar and from Jameson coke-oven tar, shows
as far as such results can do, the similarity of these crude phenols as
well as their chief difference, namely, that whereas the blast-furnace
product abounds in the phenols boiling between 200° and 230°, that
from the Jameson coke-oven tar is rather deficient in them. The
relative proportions, however, of the different phenols in the fraction
distilling between 230° and 250° are very similar in the two varieties.

CONSTITUENTS OF BLAST-FURNACE TAR. 21

Tuble comparing the Results of Fractionating the Crude Phenols from
Blast-furnace and Jameson Ooke-oven Tars, the Water being deducted.

Phenols from blast-
furnace tar,
per cent.

Phenols from Jameson coke-oven
tar, per cent.

Temperature
centigrade.

Below 160... ceseces
ea <a been
230 .. 06
230 ..

Below
e@eeeeeeeee 5°5 228°

cocecccces 10°O 385
250

9°6 270

cooccese. 5°6 800

” 360
360° to coking point .
Coke and loss........

PORN ODWNTOR ER Oe

Pitch (residue): 26 °6
Loss . .ccoe 14°8

DHOOCHOVINE AOA

Ble
oO

100°0

When similar methods of extraction are employed for oils distilling
between the same temperatures, from—

1. Ordinary Lancashire gas-retort coal-tar (“carbolic oils”), the
yield is about 5 per cent. by volume of good crude phenols, containing
65 per cent. (vol.) of a carbolic acid, sufficiently pure to crystal-
lise at ordinary temperatures with ease.

2. The blast-furnace tar, the yield was 17°5 per cent. by volume of
phenols containing the small quantity of phenol, C,H,-OH, indicated
in the table just given, and in the results recorded before the table.
(By exhaustive treatment with alkali 23:1 per cent. was extracted.)

3. The Jameson coke-oven tar-oils, the yield was about 5 per cent.,
and the amount of phenol, C,H;-OH, was extremely small, consider-
ably smaller than that obtained from the blast-furnace tar. (By
‘xhaustive treatment with alkali, about 8 per cent. could be extracted.)

It will be seen how far these results, in the case of the blast-

urnace tar, coincide with what we might expect, assuming the truth
vf Schulze’s theory, to the effect that in the formation of aromatic
hydrocarbons by the destructive distillation of coal, phenols are
primarily formed, and these subsequently suffer disruption at the
higher temperatures, yielding water and hydrocarbons of the aromatic
feries. The Jameson product is scarcely a coal-tar, and may be
tegarded as a still nearer approach to the shale oils of the parafiin
bhale distiller.

22 SMITH, COUTTS, AND BROTHERS: PHENOL

We should be very slow, however, to decide that the aromatic
hydrocarbons are always formed in this way during the destructive
distillation of coal, though strongly inclined to believe that phenols
are more easily formed first and at lower temperatures than the
aromatic hydrocarbons, and also that such phenols by decomposition
at the higher temperatures do give rise to a certain proportion of the
hydrocaibons. Still, we think it is quite possible that this may be
entirely in accordance with facts, and yet that the theories of O.
Jacobsen (Ber., 1877, 853), Berthelot (Compt. rend., 62, 905—947),
and Anschiitz (Ber., 1878, 1215), as to the synthetic formation of
hydrocarbons of the benzene series, and of naphthalene, phenanthrene,
and anthracene, may also hold good under special circumstances and
conditions. It is needless to point out that were Schulze’s theory
the only solution of the problem, it would be difficult to account for the
amount: of benzene in gas-tars, seeing that amongst the phenols so
abundantly present in low-temperature coal-tars and oils referred to
in this paper, phenol itself is so scantily present. But it will now be
necessary: to proceed with the description of the results we obtained
with the other higher boiling phenols in the group of fractions (A)
boiling below 230°.

The portion passing over between 210° and 225°, one of the largest
fractions, ‘was distilled over hot zinc-dust or passed over hot iron-
borings, the product obtained was distilled, and the distillate shaken
with caustic soda solution, to remove any unaltered phenols. The
insoluble upper layer, smelling like the ordinary “solvent naphtha”
of coal-tar, was then dried over calcium chloride and carefully frac-
tionated. It distilled between 115° and 150°, and apparently, therefore,
consisted of a mixture of xylenes with a little toluene, irdicating tkat
the material operated on was probably a mixture of xylenols with a
little cresol. The result of the fractionation showed that the chief
portion of the hydrocarbons obtained distilled between 135° and 145°.
The fraction which boiled constantly at 135—145°, and smelt
exactly like xylene, was analysed, and the following numbers were

obtained :—
II. Calculated for xylene, CsH,o,

p per cent. per cent.
Carbon : 90°51 90°56
Hydrogen... ; 9°22 9°43

A vapour-density determination (Victor Meyer’s method) gave 3°57
as the result, the calculated vapour-density of xylene being 3°67.
The material passed over the hot zinc-dust consisted of a propor-
tionate mixture of the fractions distilling between 205° and 230°.
One of the largest of these fractions distilled pretty constantly
between 205° and 215°. A portion of this shaken with water gave a

CONSTITUENTS OF BLAST-FURNACE TAR. 23

blue colour on addition of a drop of ferric chloride, and an alcoholic
solution of the phenol gave with ferric chloride a dark green tint,
turning blue on addition of water; unsymmetrical metaxylenol
[1:3:4] boils at 211°5°, and gives the same reactions with ferric
chloride. A specimen bought as pure metaxylenol from the firm of
Langfeld and Reuter of Rostock, gave the above reactions, and on
distilling a portion it was found to distil pretty much as our product
did, between 205° and 215°. Seeing that this specimen was prepared
from pure materials designed to yield the xylenol in question, it may
be considered as sufficiently proved that the substance we examined
consisted of metaxylenol [1 : 3: 4]; and seeing that it composed the
chief part of our phenol-product boiling between 205° and 230°, and
that the xylenes of gas-tar consist chiefly of metaxylene, we think
there is sufficient coincidence between these results and K. KE. Schulze’s
theory, to be of interest, and tu render them worthy of record.

Examination of the Constituents of Group B.

The fractions distilling between 230° and 3(0° were redistilled
several times, but no very definite or constant boiling points were
attained. A mixture of them was taken and distilled over hot zinc-
dust (fine iron-borings were found to answer quite as well), when a
product smelling like coal-tar naphtha was obtained. The portion of
this boiling below 180° was washed with caustic soda and water,
dried over fused calcium chloride, and then fractioned. The portion
of this boiling below 180° was washed with caustic soda and water,
dried over fused calcium chloride, and again fractioned; the portion
boiling pretty constantly between 150° and 175° was taken, but there
was too small a quantity of it to admit of making both a combustion
and a vapour-density determination ; the odour bore a strong resembl-
ance to that of pseudocumene. A vapour-density determination by
Victor Meyer’s method gave 4°01, whilst the calculated vapour-density
for trimethylbenzene is 4°15. Hence the chief phenol constituent pre-
sent in the fraction named seems to be pseudocumenol. Mesitylol was
not likely to have been present at all, at all events in more than traces,
since its boiling point is 240°, and the specimen distilled from zinc-
dust only commenced to distil at 230°. Pseudocumene, boiling at
240°, would be of course included in this, and thus the trimethyl-
benzene obtained, as before stated, might be expected to consist of
pseudocumene with but traces of mesitylene.

The portion of the reduced product boiling above 180°, was now
taken, and was found to have partially solidified to a semi-crystalline
mass. The liquid portion was poured off from this, and the crystal-
line solid residue was washed with caustic soda and then with sul-

24 PHENOL CONSTITUENTS OF BLAST-FURNACE TAR.

phuric acid. It was washed again, dried, and finally distilled. It
passed over almost entirely between 205° and 220°, whilst the fraction
distilling between 210° and 220°, solidified completely to a white
crystalline mass, which readily sublimed on application of gentle
heat, yielding beautiful crystalline plates, melting between 70° and 80°.
The odour of the substance, and its ready response to Vohl’s cha-
racteristic reaction, as well as its melting and boiling points, proved
it to be naphthalene, and the amount obtained indicates the presence
of naphthol* in considerable quantity in'the phenols boiling between
220° and 300° from the blast-furnace tar.

Examination of the Constituents of Group C. (Boiling above 300°.)

To remove all traces of hydrocarbons from these fractions, the
phenols were each redissolved in strong caustic soda solution, the
solution then largely diluted, and afterwards repeatedly shaken with
toluene. After completely removing the toluene, sulphuric acid was
added, the phenols liberated, collected, dried, and redistilled.

On passing a mixture of these higher boiling fractions twice over
red hot zinc-dust, a product was obtained yielding a fair amount of
naphthalene, again testifying to the amount of naphthols in the tar.
The higher boiling portion of the hydrocarbons consisted of a
buttery-yellowish mass, in which we failed to detect any anthracene.

It is proposed to examine this portion further so as to ascertain, if
possible, what hydrocarbons do exist in it.

On distilling over red hot zinc-dust or iron-borings, it was observed
that the higher boiling group of phenols, C, gave far less deposit of
carbon than the lower one in groups A and B.

So far then, we have proved that in the phenols obtained so abun-
dantly from blast-furnace tars (the same remarks will probably be
true of the phenols of the Aitken and Jameson coke-oven tars),
there are present—

Ordinary phenol.... C.,H;-OH.
Cresols C;H,(CH;)-OH (chiefly metacresol, 1, 3).
Xylenols C.H;(CH;).-OH (metaxylenol, 1, 3, 4).
Trimethylbenzene-

phenols C;H.(CH;);-OH (pseudocumenol).
Naphthols C,,H,OH.

The redistilled portions of the phenols boiling above- 350° were
tested to discover whether any azo-colours of acceptable shades could
be obtained from them. For this purpose, sulphanilic acid was di-
azotised, and the diazobenzenesulphonic acid treated with small

* See also Ber., 16, Ref., 150; and Aan., 227, 143 (K. E. Schulze).

GLADSTUNE AND TRIBE: ALUMINIUM ALCOHOLS. 25

quantities of the phenols in alkaline solution ; brown and red dyes
were obtained, but none of them were of inviting appearance.

It is quite possible, however, that trials of this kind might lead to
unsatisfactory results, from the fact that none of the phenols tested
could be other than mixed products.

In conclusion, we beg to state that we are far from asserting that
nothing but phenols (the homologues of C,H;-OH) are present in the
oils examined, and it is quite possible, for example, that phenol-ethers
may also be found in them. It is certaimthat some compounds not
found in ordinary gas-tar phenols similarly extracted, are present,
for whereas K. E. Sehulze (loc. cit.), in his examination of the latter
was able to isolate from the portions boiling between 230° and 300°,
both a- and A-naphthols, we found it quite impossible to obtain them
in the crystalline form from our blast-furnace tar-phenols when
following Schulze’s method. Evidently some oily or resinous
phenoloid bodies obstinately adhered: to and.accompanied the naph-
thols and prevented crystallisation.

IV.—Alumintum Aleshols. (Part ITI.) Aluminium Orthocresylate and
its Products of Decomposition by Heat.

By J. H. Guapsronz, Ph.D., F.R.S., and Atrrep Trise, F.L.C.,
Lecturer on Chemistry in Dulwich College.

Ten years ago (Chem. Soc. J., 1875, 822) we described a reacticn,
which was subsequently extended, and named “The Alumininm
Iodine Reaction” (Proc. Roy. Soc., 30, 546). By means of this a
number of aluminium-derivatives of the alcohols of the series
C,Hon4:°OH, and C,,H.»_;*OH have been prepared.

In a paper published in the Chemical Society’s Transactions, 1582,
p. 5, we described the products of the action of heat on several of
these aluminium alcohols, and more especially on those obtained from
aluminium phenylates and para- and meta-cresylates.. The general
effect of heat on these aluminium compounds is to produce alumina,
more or less of the original alcohol, and the corresponding ether, and
sometimes a hydrocarbon; but in the cases of the para- and meta-
cresylates, they each furnished, in addition, a beautiful pearly com-
pound, crystallising in hexagonal plates of the formula C,;H,,O.
These compounds were shown to be isomeric, and they were named
provisionally para- and meta-cresy] ketones,

26 GLADSTONE AND TRIBE: ALUMINIUM ALCOHOLS.

In the present paper, we describe the preparation of orthocresylate
of aluminium, and the action of heat upon it, thus completing the
study of the destructive distillation of the cresylic aluminium
alcohols.

Aluminium Orthocresylate.

The orthocresol employed in this reaction was obtained from
Messrs. Kahlbaum. It melted at 31—32°, and boiled at 186°.
Kekulé gives 31—31'5° as the melting point, and 185—186° as the
boiling point of orthocresol.

Aluminium has no action on orthocresol at ordinary temperatures,
and the action is hardly perceptible for a minute or two even at the
boiling point of the compound; but after this the action visibly
increases, and proceeds in an increasing proportion. As the action
progresses, the aluminium foil becomes studded with black specks and
patches—doubtless from the uncovering of certain electronegative
impurities which, being in juxtaposition with the active metal, probably
accounts for the acceleration in the rate of chemical change already
referred to. If a few fragments of iodine are dissolved in the cresol
when near its boiling point, and aluminium is added, the decompo-
sition takes place at once and proceeds rapidly.

In order to completely substitute aluminium for the basic hydrogen
in orthocresol, it was heated with an excess of thin aluminium foil
until action ceased. The product, while still fluid, was strained
through fine wire gauze in order to separate particles of the metal
and impurities. On cooling, it solidified to a black vitreous mass, and
gave on analysis 8°56 per cent. of aluminium. The percentage of
this metal in a compound of the formula Al,(C;H;O), should be 7°89.
The following equation would appear to represent the action :—

6C,H,OH + 2Al = Al,(C,;H,O), + 3H:.

The aluminium orthocresylate readily dissolved in benzene, giving a
dark-coloured liquid of a greenish tinge. This colour, and the black
colour of the substance itself, we have reason to think, are due to a
small quantity of some foreign substance. When the benzene solu-
tion is exposed to the air, aluminium hydrate separates, probably from
the decomposition of the cresylate by atmospheric moisture. Water
and alcohol both decompose it rapidly, aluminium hydrate being
formed, and probably the criginal cresol.

Action of Heat.

868 grams of aluminium orthocresylate, prepared as described, were
heated in a flask fitted with a wide, bent tube. It quickly melted, and

GLADSTONE AND TRIBE: ALUMINIUM ALCOHOLS. 27

at a high temperature underwent decomposition, the products being
alumina, some carbonaceous and tarry matter, and a dark-brown,
viscid distillate. This liquid on cooling was found to contain some
aluminium hydrates, but little or no solid organic compound, thus
differing from the corresponding volatile products of the destructive
distillation of the para- and meta-cresylates. The distillate was
divided by fractionation into three portions—(a) boiling between 140°
and 200° ; (b) boiling between 200° and 300°; and (c) boiling above
300°.

Fraction a.—This portion was a liquid of a light yellow colour
weighing 65 grams. It readily dissolved for the most part in a
solution of potassium hydrate, from which it again separated on
neutralisation with hydrochloric acid. The liquid thus separated was
dried over calcium chloride. It boiled at 185—186°; its sp. gr. at
182° was 1:049 ; its refractive index for the line A was 1°5373, and for
the line H, 1°5851. On heating with aluminium foil and a fragment
of iodine, it quickly underwent decomposition with evolution of
hydrogen. This fraction, therefore, consisted almost wholly of the
original alcohol reconstituted.

Fraction -b.—This portion weighed 200 grams, and was of a yellow-
ish-brown colour. It was well shaken with a solution of potassium
hydrate, and the residue washed with water and dried over calcium
chloride. The liquid obtained was then fractionated some ten times,
rejecting at each distillation the small quantities boiling respectively
below 200° and above 300°. The product was further fractionated
about 20 times, gradually eliminating portions having the greatest
range of temperature. In this way a fraction was ultimately
obtained (about one-eighth of the whole) which may be supposed to
be a fairly pure specimen of the compound, of which (6) in a great
part consists.

The substance thus isolated was a colourless, moderately mobile
liquid, which became slightly yellow on exposure to light. Its boiling
point was 272—-278°, and at 243° its sp. gr. was 1°047; its refrac-
tive index for the line A was 1°5638, and for the line H 1°6202. On
combustion with oxide of copper and oxygen :—

I. 0°238 gram gave 0°7456 gram CO, and 0°1502 gram H,0.
II. 0270 - 0°8449 - 0°1698 -

The results expressed in parts per 100 give—

On the determination of its vapour-density, 0°1773 gram gave vapour

28 GLADSTONE AND TRIBE: ALUMINIUM ALCOHOLS.

= in volume to 20c.c. (corr.). The numbers give 197°8 for the
molecular weight of the substance. Its most probable molecular
formula is therefore C,,H,,O.

Cale. for 100 parts. Found. Mean.
85°38
7°00
7°62 (diff. )

100-00 10u°00

The odour of this substance resembles that of phenylic ether—
numely, that of the geranium leaf,* but fainter, and is similar to the
odour of para- and meta-cresyl ethers. We think we are therefore
justified in concluding that the substance is also a cresyl ether. It did
not crystallise when surrounded by a freezing mixture, from which it
would appear not to be the para-modification already described. It
seems also to differ in boiling point and sp. gr. from the metacresylic
ether already described in the same way, and to about the same extent
as the ortho- differs from the meta-cresylic alcohol. We conclude,
therefore, from this, as well as from the origin of the compound, that
it is orthocresylic ether.

Boiling Sp. gr. at Specific Specific

Cresylic aleohols. point. 18 °2°. refraction. | dispersion.

Ortho ccscccccccsccces 185°C. 1°049 0°5122 0 -0456
196—202° C. “04: 0°5116 00454

Boiling p. gr. Specific Specific

Cresylic ethers. ; 2. 2ageee ,
. point. . refraction. | dispersion.

Ortho coccccccccccces.| 979-878" 04 0°5385 | 0°0266(F-A)
Behe ccccccccccccccccs| S06—O08 1028 0°5386 | 0°0266(F-A)

It will be observed that each pair of isomeric compounds is practi-
cally identical in specific refraction and dispersion, which gives great
confidence in the purity of all four specimens,

Fraction c.—This weighed 185 grams, and consisted of a dark-
brown, highly viscous fluid. The portions of b boiling above 278°

* This resemblance was so marked, that we examined a specimen of pure essential
oil of geranium, in the hopes of finding that it contained our cresyl ether; but it
dil not. We were equally unsuccessful with the compound known as “ Indian
geranium,”

GLADSTONE AND TRIBE: ALUMINIUM ALCOHOLS. 29

were added to this fraction, and the whole distilled several times;
each time collecting, separately, small quantities of liquid which
passed over below 280°, and rejecting other small quantities of a tarry
nou-distillable substance left in the retort. The distillate ultimately
obtained was lighter in colour and much less viscous than the original
fluid, and on cooling, and still more on standing for 24—48 hours,
deposited small quantities of a yellow crystalline substance. In
order to separate this solid from the liquid, rather more than an
equal volume of alcohol was added, and then enough ether to dissolve
the liquid. This mixture wk«t subjected to a freezing mixture of
ice and salt, yielded a small c.iartity of the crystalline substance,
which was removed by ‘filtrat.cn, The alcohol and ether were
next separated from the liquia substance by distillation, and the
residuum distilled, rejecting, as before, portions boiling below 28°
and the tarry non-distillable residue. The distillate obtained was again
mixed with alcohvl and ether, and subjected to a freezing mixture
when another portion of the crystalline compound separated. These
operations were repeated some 15—20 times, as Jong as any quantity
of the solid could be recovered. The total amount isolated weighed
a little less than’2 grams.

In order to purify this substance, it was dissolved in boiling alcohol
and recrystallised therefrom some four or five times. It was then
slowly distilled, and the distillate again crystallised from alcohul. The
body was now lemon-yellow in colour, and consisted of fragments of
very thin plates. On combustion with oxide of copper and oxygen, it
gave numbers agreeing’ with the formula C,;H,,0.

In 100 parts.
A.

~,
Calculated.

85°71
6°66
7-63 (diff.)

100°00

The experience gained in the purification of some of the compounds
described in the second part of this research led us to suspect that the
lemon-yellow colour of the substance above referred to was occasioned
by its admixture with a small quantity of colouring matter. With
the object of separating this impurity, the substance was slowly sub-
limed, when brilliant plates were obtained, but of no very definite
form. These were dissolved in hot alcohol, and the solution on cooling
deposited diamond- and star-shaped plates free from colour. The
quantity of the substance thus purified was too small to allow cf
further quantitative examination. We have little doubt, however,

30 BRIERLEY ON SOME NEW VANADIUM COMPOUNDS.

but that its composition is expressed by the number given above, and
that it has the molecular formula C,;H,,0. Whether the body is
identical with either of the so-called ketones obtained from the
aluminium meta- and para-cresylates, or is an isomeride bearing the
same relation to orthocresol as these compounds bear to their respec-
tive alcohols, we have no evidence, physical or chemical, which will
enable us to decide, as we attach little or-no importance to-the difference
in crystalline form.

It appears from this investigation that the action of heat on the
aluminium orthocresylate is generally similar to the action of heat on
the para- and meta-modifications. Each of these compounds has
yielded its hydrogen-derivative, or alcohol, the corresponding ether,
and a crystalline sublimable compound of the formula C,;H,,0,
together with others not isolated. There are, however, one or two
points of difference of some interest. Firstly, the different aluminium
compounds have yielded very different quantities of distillate, and,
still more, very different quantities of the so-called ketones. This is
shown in the following table of results :—

Distillate from Solid from

1000 parts. 1000 parts.
Aluminium paracresylate..... 609 106
metacresylate.... 649 20

9
orthocresylate.... 475 2°5

Another point which may perhaps account for the different amounts
of distillate, is the different temperatures required to effect destruc-
tive distillation of the several aluminium cresylates. This was
especially noticed in the case of the para- and ortho-modifications.
Besides probably accounting for the difference in the amount of the
product, this observation is of interest as showing that a difference
exists in the stability of these aluminium isomerides.

V.—-On some New Vanadium Compounds.

By J. T. Brizrtey, Dalton Scholar, Owens.College.

THESE compounds are formed by the-following remarkable reaction.
If a blue solution of hypovanadic sulphate be mixed with a colourless
solution of an alkaline metavanadate, a dark-green liquid will be pro-
duced, and if to this a slight excess of caustic soda is added, the colour
of the solution quickly changes to a deep black. From this dark+

BRIERLEY ON SOME NEW VANADIUM COMPOUNDS. 31

coloured solution, well-defined crystalline salts having a purple or dark-
green colour and metallic lustre, can be obtained, in which the condition
of oxidation of the metal is intermediate between the tetroxide, V,O,,
and the pentoxide V,0;. I have succeeded. in preparing five distinct

numbers of this group of salts, viz. :—
1. A soluble sodium salt having the composition—

2V.0,,V,0;,2Na,0 + 13H,0.

. A soluble potassium salt—
2V,0,,V,0;,2K,0 + 6H,0.
. An insoluble potassium salt—
2V.0,,4V20;,5K,0 + H,0.
. A soluble ammonium salt—

2V20,,2V20;,(NH,),0 + 14H,0.

. An insoluble ammonium salt—

2V.0,,4V20;,3(NH,).0 + 6H,0.

Intermediate Oxides.

Professor Roscoe observed long ago that when black vanadium

trioxide, V,0 3, is exposed to the air for many months, it absorbs
oxygen and moisture from the air, its colour changing to a pale
grass-green. On analysing a sample of such an oxide which had been
freely exposed to the air for a considerable length of time, I found
that its composition, as regards its degree of oxidation, corresponds
with that of the insoluble potassium and ammonium salts, viz.,
V.0,,2V.0; + 8H,0.

Calculated.
45°49
33°19
21°32

100°00

This oxide, dissolved in dilute sulphuric acid and neutralised with
caustic potash, yields, when heated, the purple insoluble potassium
salt.

Another intermediate dark-green oxide, containing an equal number
of molecules of tetroxide and pentoxide, may be obtained by the
gentle ignition of the ammonium salt, No. 5. This oxide readily
absorbs moisture from the air, its composition then being—

3(V2Q,,V.0;) + 8H,0.

BRIERLEY ON SOME NEW VANADI@M COMPOUNDS.

Calculated.
“51°62
36°29
12-09

100-00

This intermediate oxide, V,O,, can also be obtained in solution by
adding strong sulphuric acid to a hot saturated solution of ammonium
metavanadate, and then passing sulphur dioxide through the solution
until the liquid attains a green colour. After driving off the excess
of sulphur dioxide by boiling, the solution was titrated with perman-
ganate and found to contain vanadium tetroxide and pentoxide’ in
equal molecular proportions. The green-coloured solutions thus
obtained may also be prepared by suspending freshly precipitated
hydrated vanadium tetroxide in water, adding an insufficient quantity
of dilute-sulphuric acid to dissolve it, and allowing the turbid liquid
to stand. several hours. These solutions are very unstable, gradually
changing colour from absorption of atmospheric oxygen, -passing
through all shades of colour, from dark-green to light-brownish
yellow.

Intermediate Salts.

1. Soluble Sodium Salt, 2V,0,,V,0,;2Na,.0 + 'I3H,O.—In order to
prepare this salt, 12 grams of finely powdered vanadium pentoxide
is placed in a flask, an excess of a strong solution of sulphur
dioxide poured on to it, and the solution heated to the boiling point,
small quantities of sulphuric acid being added from time to time.
The deep blue-solution thus obtained is then well boiled to expel every
trace of sulphur dioxide. Six grams of vanadium pentoxide are
separately boiled with an excess of caustic.soda until completely dis-
solved. The two solutions are now mixed hot, and caustic soda added

‘to the dark blackish-green liquid until it is slightly alkaline; after
standing for a short time, acetic acid should be added to the liquid in
very slight excess only, as an excess of caustic soda decomposes the
green salt. On filtration after boiling, the filtrate contains the new
salt together with sodium sulphate. As these two salts possess nearly
the same degree of solubility they cannot be separated by crystallisa-
tion; on adding to the cold solution a cold saturated solution of sodium
acetate, however, the new vanadium salt is precipitated in black
shining crystals. These are deposited in large quantity and of con-
siderable size if the liquid is heated after addition of the acetate, and
then allowed to cool. The crystals should be washed with a saturated
solution of sodium acetate until’ the washings are free from sulphuric
acid, and the adhering acetate removed by washing with dilute alcohol

BRIERLEY ON SOME NEW VANADIUM COMPOUNDS. 33

until the latter no longer yields an acid distillate. Finally the crystals
are dried over calcium chloride. The sodium salt thus obtained con-
sists of black shining plates which, under the microscope, are seen to
consist of groups of short hexagonal prisms, and have a sp. gr. of
1327 at 15° compared with water at 15°. A solution of the pure salt,
on slow crystallisation, yields hexagonal plates of considerable size,
having the form shown in the figure, and an angle between the faces
of the pyramid and prism of 133° 45’. The salt is readily soluble in
water, yielding a dark blackish-green solution which, on acidification
with a few drops of sulphuric acid, instantly changes to a light grass-
green, whilst alkalis restore the black colour, unless an excess be
added, when the compound is decomposed with formation of a red-
brown solution. The sodium salt is insoluble in strong saline solu-
tions—especially in those of the acetates in the cold; if boiled with
potassium or ammonium acetate, purple insoluble salts of these metals
are formed.

The quantity of the two oxides of vanadium present was ascertained
by titration with a standard permanganate solution, care being taken,
in the determination of the tetroxide, to use cold solutions, to displace
all air by carbon dioxide, and to use well boiled water for dissolving
the salt. The total vanadium was estimated by the same method after
complete reduction with sulphur dioxide. In order to estimate the
alkaline metal, the vanadium was precipitated as lead salt by basic
lead acetate, the excess of lead removed by sulphuretted hydrogen,
and the sodium weighed as sulphate.

For the purpose of determining the water of crystallisation, the
salt was heated to redness in a combustion-tube in a current of dry
air, and the water collected in a weighed calcium chloride tube.

The following are the results obtained :—

VOL. XLIX.

34 BRIERLEY ON SOME NEW VANADIUM COMPOUNDS.

Found. Calculated.
r a * ~ 2V,0,,V,0;,2Na,0 +
lst preparation. 2nd preparation. 13H,0.
V.0,.... 37°67 36°80 38°12
V,05.... 21°19 21°00 20°88
H,O .... 26°05 25°34 26°80
Na,O.... 1415 13°88 14°20

99°06 97°02 100-00

2. Soluble Potassium Salt, 2V20,,V,0;,2K,0 + 6H,O.—In order to
prepare this salt, 20 grams of vanadium pentoxide were reduced by
sulphur dioxide, and 10 grams also were converted into potassium
metavanadate, the filtered solutions being mixed and caustic potash
added to slight alkalinity. The resulting greenish-black liquid was
then heated to the boiling point, filtered, mixed with a cold saturated
solution of potassium acetate, and the mixed liquids evaporated on
the water-bath to a small bulk. On cooling and standing, a mass of
small dark crystals separated which, under the microscope, were seen
to consist of imperfectly formed octohedrons having a purple colour,
mixed with other dark-green four-sided crystals. The mixed crys-
talline mass was then well washed on a filter with hot water, which
dissolved the greenish-black crystals, leaving the purple insoluble salt
behind. The dark-green filtrate was, as before, mixed with a concen-
trated solution of potassium acetate, heated until the separated crys-
tals redissolved, and the solution left to cool. On standing, fine
greenish-black crystals were deposited from which all trace of potas-
sium sulphate was removed by continued washing with weak alcohol.
This compound has a sp. gr. of 1°389 at 15°, compared with water
at 15°.

Analysis of this salt gave the following :—

Found.
A Calculated.

"Ast 2nd preparation. 2V;0,,V20;,2K,0
preparation. -— + 6H,O.
V0, .... 42°02 40°83 , 41°04
V0, .... 23°54 23°90 2° 22°48
HO .... Iti 12°82 ; 13°31
K.0 .... 21°04 21°21 ‘ 23°17

97°71 98°76 ; 100-00

Preparation No. 1 contains 5 mols. H,0.

3. Insoluble Potassium Salt, 2V.0,,4V.0,5K,0 + H,0.—The
purple insoluble salt above mentioned can be easily obtained pure
from the motber-liquors of the sodium salt. This dark-green solu-

BRIERLEY ON SOME NEW VANADIUM COMPOUNDS. 35

tion containing sodium acetate and sulphate is heated and filtered
hot, to the hot liquid some solid potassium acetate is added, and the
whole is well boiled. Small crystals having a metallic lustre soon
form on the surface of the liquid, and, if the boiling be continued, the
whole of the soluble sodium salt may thus be converted into the
insoluble potassium compound. The sp. gr. of this salt is 1'213 at
15° compared with water at 15°. In order to determine the potas-
sium in this insoluble salt, it was oxidised with strong nitric acid,
and the pentoxide converted into the ammonium salt; this was then
precipitated with lead acetate as before, and the alkali weighed as
sulphate.

Found.

r ~ Calculated.

lst preparation. 2nd preparation. 2V,0,,4V.0;,5K,0 + H,0.
V.0,.... 20°75 22°81 21°48
V.0;.... 47°47 45°98 47°05
K,0 .... 28°24 29°82 30°32
ae 2°26 1:17 115

98°72 99°78 100-00

The amount of water in Preparation No. 1 corresponds with 2 mols
H,0, whilst that in No. 2 agrees with 1 mol. H,0.

4. Soluble Ammonium Salt, 2V,0,,2V205,(NH,),0 + 14H,0.—This
salt is formed by boiling vanadium pentoxide with ammonia until
solution takes place, reducing two-thirds of the resulting solution to
V,0,, mixing this with the unreduced third, and adding ammonia to
slightly alkaline reaction. A dark-green solution of the above
ammonium salt is thus formed; this cannot be boiled, as the salt is
thereby converted into the insolublecompound. By adding acetic acid
to the solution until faintly acid, and then alcohol, and allowing the
mixture to stand after well shaking, a black crystalline precipitate
separates. The supernatant liquid is decanted, and the precipitate
thrown on a filter, and washed with a mixture of equal volumes
of alcohol and water until free from sulphuric acid. The salt thus
obtained is then dried over calcium chloride. It consists of greenish-
black crystals which are very unstable, readily absorbing oxygen from
the air. In order to determine the ammonia, the salt placed in a
platinum boat was heated in a combustion-tube, and the ammonia
collected and weighed as platinochloride. Analysis gave :—

36 BRIERLEY ON SOME NEW VANADIUM COMPOUNDS.

Found.
1st preparation. Calculated.

V:0, .... 33:20 33-23
V:0; .-.. 35°62 36°43
(NH,),0. 5°67 5°19
H,O*.... 25°51 25°15

100°00 100-00
* Water by difference.

5. Insoluble Ammonium Salt, 2V,0,,4V.0;,3(NH,).0 + 6H,0.—
The foregoing soluble salt is readily converted into this insoluble
compound by adding ammonium chloride to its solution rendered
alkaline by ammonia, and gently heating the liquid for some time.
Small purple crystals having a metallic lustre separate out—first on
the sides of the vessel, and afterwards on the surface of the hot
liquid. It is not necessary for this purpose to prepare the pure
soluble ammonium salt, as the dark-green liquid obtained by mixing
equal parts of solutions of V.0, and of ammonium metavanadate
readily yields this salt when rendered alkaline with ammonia and
heated. The crystalline precipitate is well washed with hot water

until it is seen under the microscope to be perfectly homogeneous.
It crystallises in small purple octagonal plates, of a metallic lustre,
and yields a brown powder. Its sp. gr. at 15° is 1335 compared

with water at 15°. Analysis gave :—

Found.
A ~ Calculated.
lst 2nd preparation. 2V,0,,4V,0;,3(NH,).0
preparation. ~————*~——— + 6H,O.
V.O, .... 25°64 24°77 = =25°15 25°09
V.0, .... 55°48 50°29 = 55°42 55°01
(NH,).0.. 11°68 12°40 11°51 11°76
H,O* .... 7°20 7°54 7°92 814

10000 100:00 100-00 100:00
* Water by difference.

In concluding my paper, I must thank Mr. Harden, of Owens
College, for kindly confirming several of my results.

ft =? 2. a aa

SS em GY VS Vw oF ee

VI.—On the Vapour-pressures of Mercury.
By Wittiam Ramsay, Ph.D., and Sypney Youne, D.Sc.

In a recent paper in the Transactions (1885, 640—657), we described
a method by which constant known temperatures could be maintained
with great exactitude. In the tables which conclude that paper,
Regnault’s determinations of the vapour-pressures of mercury have
been accepted as correct; on revision, however, they appear to be
by no means so trustworthy as most of his work. In vol. 26 of the
Mémoires de l’ Académie Regnault quotes results described in vol. 21,
p. 502, and uses these results for determining the formula devised
by him to express the relation between temperature and pressure at
low temperatures. On reference to vol. 21, it is evident from
Regnault’s own remarks, that he placed little confidence in the
accuracy of his results. His words are as follows :—‘ Ces deux séries
d’expériences différent notablement quand on compare les valeurs
relatives qu’elles donnent pour les forces élastiques de la vapeur mer-
curielle; mais les différences absolues que 1’on trouve entre les forces
élastiques sont réelement trés-petites, et de l’ordre de l’incertitude des
observations. Les expériences précédentes suffisent neanmoins pour
montrer que la tension de la vapeur de mercure 4 100° est d’environ
0°5 mm. ; et qu’ la température de 50°, elle s’éléve 4 peine 4 0'1 mm.
Elle est donc 4 peu prés négligeable au dessous de 50°.” These words
refer to temperatures below 100°. In a footnote he proceeds—“ J’ai
fait quelques déterminations de la tension de vapeur du mercure a
des températures plus élevées, en chauffant le ballon dans un bain
d’huile. Vers 200° les expériences sont devenues évidemment fautives,
par suite de la distillation du mercure. Quoiqu’il en soit, voici les
nombres que j’ai obtenu: «1 convient de ne les regarder que comme des
approximations.” (The italics are ours.) His numbers then follow.
They are the results of eleven readings, and these are all which he
gives at temperatures below 250°. He also refers in vol. 26 to four
determinations given in vol. 21, p. 230, made at the boiling point of
mercury under atmospheric pressure, for the purpose of comparing
his mercurial and air thermometers. His method in this instance
differed essentially from the other methods employed; it consisted in
heating an air reservoir terminating in a capillary tube in the vapour
of mercury boiling under atmospheric pressure, sealing the tube, and
when cold, breaking the point off under mercury, and measuring the
contraction of the air. These four experiments gave fairly concordant
results, but owing to an evident misquotation in vol. 26, it is not

38 RAMSAY AND YOUNG ON THE

quite certain to what pressures his symbols h and hy refer, as from h
the mean of the four determinations would give the pressure 767°43,
while from h, 765'29 is the mean. But as the temperatures them-
selves differ by nearly 2° at the same pressure, this slip is an unimpor-
tant one. At higher pressures, Regnault employed a large iron still,
containing 50 kilos. of mercury, into which dipped an iron tube con-
taining the air thermometer. By applying pressure from an air
reservoir, the boiling point of the mercury was regulated and
observed. He states that the temperature was observed as soon as it
had become constant, and the observations were repeated until it
began to rise, by which time the greater part of the mercury had
distilled over. He directs attention to the fact that, using this
method, the temperatures are those of the boiling liquid, and not of
the vapour. During distillation, the noise of the bumping was com-
pared to blows of a hammer on an anvil. That these results cannot
be relied on is shown by those referring to the atmospheric pressure,
where a fall of 74 mm. in pressure corresponds to a rise of tempe-
rature of 1°86°; and the mean temperature (about 355°4°) by no means
corresponds with that obtained by the method previously described
(358°5°). It is obvious that good results cannot be anticipated under
such conditions.

The importance of a correct knowledge of the vapour-pressures of
mercury is evident, when it is considered that they enter into all
calculations of the determinations of vapour-pressures of liquids by
the usual method at high temperatures; and that they must be
allow ed for in determining vapour-densities by Hofmann’s apparatus.
Moreover, it is probable that thermometers based on a knowledge of
the vapour-pressure of mercury will become available for accurate
determinations of high temperatures.

Owing to the discovery of certain relations between the vapour-
pressures of different substances, which are described in the Philo-
sophical Magazine for December, 1885, and January, 1886, we have
been able, with small expenditure of time and labour, to obtain results
which we think will be generally accepted as correct. Before describing
the experiments, a short description of the generalisations may render
the process intelligible. Although some of these apparently have
little connection with the matter in question, yet from thermo-
dynamical considerations, they are so interlaced, that we deem it
advisable to state them all. For detailed description and proof of
the following statements, we would refer to the articles already
mentioned.

1. The amount of heat required to produce unit increase of volume
in the passage from the liquid to the gaseous state at the boiling point
under normal pressure, is approximately constant for all substances.

VAPOUR-PRESSURES OF MERCURY. 39

2. If the amounts of heat required to produce unit increase of
volume in the passage from the liquid to the gaseous state be com-
pared at different pressures for any two bodies, then the ratio of this
amount at the boiling point under a pressure p, to the amount at
another pressure, #2, is approximately constant.

3. It appears probable, from the few data at present available, that
if a diagram be constructed in which the ratios at the same pressure
between the heats of vaporisation of two liquids, at various pressures,
the same for both, form the abscisse, and the absolute temperatures
of one of the two liquids, corresponding to those vapour-pressures,
form the ordinates, then the points representing the relations between
these ratios and the absolute temperatures will lie in a straight line.

4. The products of the numbers representing the absolute tempe-
rature into the increase of pressure per unit rise of temperature at
those temperatures, are approximately the same for all substances at
the same vapour-pressure; but the differences are real and are not
due to error of experiment or calculation.

5. The rate of increase of this product with rise of pressure is very
nearly constant for all bodies.

6. The deviation from constancy presents the following relations :—
If a diagram be constructed in which the ratios at definite pressures
of the product referred to in (4) for any two substances be made the
abscisse, and the absolute temperatures of one of the bodies at corre-
sponding vapour-pressures be made the ordinates, then the points
representing the relations of these two quantities will fall in a
straight line.

7. A relation exists between the ratios of the absolute temperatures
of all bodies, whether solid or liquid, whether stable or dissociable,
which may be expressed in the case of any two bodies by the equation
R' = R + c(t’ — #); where R is the ratio of the absolute tempera-
tures of the two bodies, corresponding to any vapour-pressure, the
same for both ; R’, the ratio at any other pressure, again the same for
both; c, a constant, which may equal 0 or a small plus or minus
number ; and ¢’ and ¢ the temperatures (absolute or centigrade) of one
of the bodies corresponding to the two vapour-pressures.

When c = 0, R’ = R, or the ratio of the absolute temperatures is a
constant at all pressures; and when c > 0, orc < 0, its value may
readily be determined either by calculation, or graphically by repre-
senting the (absolute) temperatures of one of the two bodies
as ordinates, and the ratio of the absolute temperatures at pressures
corresponding to the (absolute) temperatures of that body as abscisse.
It is found in all cases that points representing the relation of the
ratio of the absolute temperatures of the two bodies to the (absolute)
temperature of one of them fall in a straight line.

40 RAMSAY .AND YOUNG ON THE

On comparing such relations for different substances of which the
vapour-pressures had been accurately determined, in twenty-three
cases, this relation was found to hold absolutely ; for on calculating
by means of this ratio the absolute temperatures of these substances
corresponding to given pressures, the differences observed between
calculated and experimental results fall well within the limits of ex-
perimental error. On using Regnault’s determinations of the vapour-
pressures of mercury, the case was different. Here a curve, instead
of a straight line, was obtained. Now it is evident that if satisfactory
proof can be given that certain temperatures, sufficiently far removed
from each other, when compared with the ratios between these
absolute temperatures of mercury and some other liquid of known
vapour-pressure, ¢.g., water, give points which fall in a straight line,
the value of ¢ is calculable, and as a consequence, the whole vapour-
pressure curve of mercury. Mercury also formed the only excep-
tion to statements (5) and (6).

The temperatures chosen were—(1l) the boiling point of methyl
salicylate under atmospheric pressure; (2) the boiling point of
bromonaphthalene under a pressure of 612°8 mm.: (3) the same
under a pressure of 756°2 mm.; (4) more reliance is to be placed on
the boiling point of mercury under atmospheric pressure, deter-
mined by Regnault, described in vol. 21 of the Mémoires, as the
method he adopted seems to deserve greater confidence than his later
method ; and the mean of his four results may be taken as fairly
correct ; (5) the boiling point of sulphur under atmospheric pressure,
determined by Regnault. For the accuracy of our knowledge of the
first three of these temperatures reference must be made to the
Transactions (1885, p. 640 et seq.).

Regnault determined the vapour-pressures of sulphur, using an
apparatus similar to that with which he determined the vapour-
pressures of mercury. He did not find it necessary to fill the still
with liquid sulphur, and the boiling took place quietly, without
bumping. He appears to place more reliance on his determinations
of the vapour-pressures of sulphur than on those of mercury ; and
this conclusion is borne out by our experience ; for on comparing the
vapour-pressures of sulphur with those of carbon disulphide, the
ratios of the absolute temperatures form with the absolute tempera-
tures of carbon disulphide, or sulphur, a straight line, when con-
sidered as described in statement (7), and the results agree remark-
ably well with statements (4), (5), and (6). It is obvious that the
temperature of the boiling point of sulphur under atmospheric pressure
really amounts to an indirect reading of Regnault’s air thermometer ;
and for the above-mentioned reasons we believe that the temperature
in this case is correct.

VAPOUR-PRESSURES OF MERCURY. 41

For the determinations of the vapour-pressure of mercury, the
process adopted in our late experiments was employed; and the
apparatus is represented in a woodcut on p. 643 of the Transactions
for 1885, modified as described on p. 651. A description is therefore
unnecessary here.

(1.) The detailed results are as follows :—

Closed Open | Vapour

limb of | limb of . | Pressure pee pressure of
apparatus. | apparatus. : —_* mercury.

703°8 704°5 ’ , ; 35°1
704°2 705°0 g ; ; 35 *2
701°1 708°0 i ; 41°05
704°3 704°1 : 2 | ; 34°2

Note.—Two gauges and barometers were used, lettered A and B respectively.
The mean reading of the two was taken.

The pressure under which the methyl salicylate boiled was
746°95 mm. (reduced to 0°), and the corresponding temperature 222°15°.
The mean observed vapour-pressure of mercury at this temperature
is accordingly 344 mm.

(2) and (3.) The detailed results of our observations with bromo-
naphthalene are given in our previous paper (Joc. cit.). The numbers
are as follows :—

Pressure of bromo- Pressure of mercury Temperature from
naphthalene vapour. vapour. Regnault’s data.

756°2 mm. 157715 mm. 280°6°
6128 ,, 12435, 270°35°.

These temperatures are calculated from Regnault’s formula for the
vapour-pressures of mercury, and were made use of in determining
the position of the curve representing the relations of temperature to
pressure for bromonaphthalene. It was noticed at the time that this
curve would not pass through both points representing the tempe-
ratures of the bromonaphthalene; and as other readings had been
taken with a mercury thermometer previously compared with an air
thermometer, confirming the lower temperature, it was judged right
to construct the curve on this basis. Reading then from the curve
we adopted, the higher temperature becomes 280°2°, instead of 280°6°.
The justice of this correction has since been confirmed by a study of
the relations between the absolute temperatures of bromonaphthalene
and water.

(4.) The relations obtained for mercury at the atmospheric pres-

42 RAMSAY. AND YOUNG ON THE

sure by Regnault have already been considered. They are as
follows :—

Temperature. H,. Ho.
358°46° 769°59 mm. 770°27 mm.
357°48 768°40_,, 766°76
359°27 768°08 ,, 766°11
358°68 763°65_,, 758°02

H, and H, are the barometric heights at the moment of closing the
air thermometer, and of breaking the point off under mercury; but
it is uncertain from the text which symbol is to be used for the one
and which for the other. The mean temperature is 358°47°; the
mean of h, is 767°43 mm.; and the mean of ho, 765°29 mm.

(5.) In order to determine the vapour-pressure of mercury at the
boiling point of sulphur, a new form of apparatus was employed,
which is shown in the accompanying figure.

The apparatus AA’ is constructed of barometer tubing of small
bore. At the end A is blown a bulb so adjusted as regards size that
the volume of the bulb is approximately equal to that of the hori-
zontal portion of the stem, and when cold, containing mercury to the
marka. Bisa vessel containing sulphur ; in one experiment this was
a flask, as shown ; in the other, a wide test-tube. On applying heat,
the sulphur vapour could be caused to completely surround the bulb
A. The mouth of the flask was loosely closed with cotton-wool, to
prevent escape of sulphur vapour, or the formation of air currents.
C is a jacketing tube in the form of a Liebig’s condenser, whereby
the horizontal portion of AA’ could be maintained at a constant
known temperature, shown by the thermometer D.

The tube AA’ was first graduated in millimetres from about a to
the end ; it was then exhausted, filled with mercury, and repeatedly —

VAPOUR-PRESSURES OF MERCURY. 43

boiled out under atmospheric pressure, until air bubbles ceased to come
off from the glass. The end A’ was next drawn out to a capillary tube,
and was exhausted repeatedly with a Sprengel’s pump, dry air, which
had stood in contact with phosphorus pentoxide for some time, being
admitted after each exhaustion. While the capillary tube was still
connected with a small drying tube of pentoxide, the tube AA’ was
placed in a vertical position in a wide tube through which a current
of water of constant known temperature ran; the capillary end of
AA’ projected just above the surface of the water. The temperature
of the water and the atmospheric pressure were then read, and the
capillary was sealed, the position of the mercury in the tube being
noted at the same moment. From these data, the volume of the air
at known temperature and pressure was subsequently ascertained.
The apparatus was then arranged as shown in the figure, and the
sulphur made to boil. The ebullition took place quietly and easily,
and thin deposit of liquid sulphur on the sides of the flask made it
certain that the vapour was not superheated. Had this not been the
case, the bulb would have been surrounded with asbestos. When the
mercury in the bulb had become hot, it was driven by the pressure of
its vapour nearly out of the bulb, and was forced along the horizontal
portion of the tube, as far as b, compressing the air. The position of
b was then read, and the temperature of the water in the jacket C
was noted. The barometric pressure was also read, so as to determine
the temperature of the sulphur vapour. When cold, the tube was
broken close to the’ bend, and calibrated by weighing with mercury.
The pressure of the mercury vapour in the bulb is equal to that of the
air in the gauge, plus the pressure of a column of mercury equal to
the vertical distance between the level of the mercury in the bulb
and that in the horizontal portion of the tube, this column being
corrected for temperature.
The detailed results are as follows :—
I, (1.) Air at atmospheric pressure :—

Pressure
Temperature. (reduced to 0°). Volume.

16°20° 743°7 mm. 0°91690 c.c.
(2.) Air at high pressure :—
Temperature. Volume.

16°60° 0°24440 c.c.

Pressure calculated from above data, 27940 mm.
(3.) Column of mercury heated to 450°, 57 mm.
~ about 100°, 51 mm.
These reduced to 0° become 52°8 mm. and 50° 1 mm. respectively.

44 RAMSAY AND YOUNG ON THE

The total pressure of the mercury vapour is therefore 2794 + 52°8 +
50°1 = 2896°9 mm.

(4.) The barometric pressure when the sulphur was boiling was
742°6 mm. (reduced to 0°), and the corresponding temperature was
447°0°.

II. (1.) Air at atmospheric pressure :—

Temperature. Pressure. Volume.

16°24° 752°3 mm. 0°53546 c.c
(2.) Air at high pressure :—

Temperature. Volume.

15°40° 0°14385 c.c.

Pressure calculated from above data, 2791‘0 mm.
(3.) Column of mercury heated to 450°, 52°5 mm.
about 100°, 66°0 mm.

” ” 9

These reduced to 0° become 48°6 mm. and 64°9 mm. respectively.
The total pressure of the mercury vapour is therefore 2791 + 48°6 +
649 = 29045 mm.

(4.) The barometric pressure when the sulphur was boiling was
754°4 mm. (reduced to 0°), and the corresponding temperature was
448°0°.

The whole of these results are shown in the following table :—

Absolute Ratio of

Temperature | Temperature temperature absolute

(centigrade). | (absolute). of water at temperature.
pressure p.

222°15° | 495-15° 34°40 mm. | 304°5° 16262
270°30 | 543-80 124°35 | g29-2 16504
280°20 | 553-20 157°15 334-2 16553
nd | hy 769°59 | 373°35
358° 46 631°46 x 570-37 a 16913
h, 768°40 | 373-80 1°6889
630 °48 ho 766°76 373°25 1°6892
; hy 768-08 | 373-29 1°6938
632 27 ho 766 *11 | 373-22 1°6941
h, 763°65 373 ‘13 1 -6930

357 °48
359 27

720-0 | 2896-9 | 415°26 1 -7338
721°0 | 2904°5 415°36 1°7359

447°0
448°0

|
|

The accompanying diagram, in which the ratios are represented as
absciss#, and the absolute temperatures of mercury in one case and
of water in the other are the ordinates, represents the results obtained.

VAPOUR-PRESSURES OF MERCURY. 45

The ratios calculated from Regnault’s determinations give the dotted
curve with absolute temperatures of water as ordinates.

It will be noticed that in each case a straight line can be drawn
through all the points. The value of c, if the temperatures of mer-
cury are chosen as ordinates, is 0°0004788; if those of water are
chosen, it is 0°0009792.

The calculations of the vapour-pressures of mercury from these two
constants give slightly different results; but the difference within the
limits of temperature given in the ensuing tables is unimportant,
although at higher temperatures it might become considerable. It
was more convenient to employ the constant derived from the abso-

46 RAMSAY AND YOUNG ON THE

inte temperatures of mercury; and we now proceed to show the
method of calculation.

A point was read from the straight line, giving the ratio at any
one temperature. The absolute temperature of water was calculated
from the ratio. The vapour-pressure of water corresponding to this
temperature is the same as that of the mercury, inasmuch as the
ratios refer to equal pressures. Thus, at an absolute temperature of
mercury of 508°, the ratio as read from the line was 1°6331. The
absolute temperature of water was therefore reat = 311°06°. The
vapour-pressure of water at 311:06°, ascertained from Regnault’s
tables is 49°466 mm., and this is therefore the vapour-pressure of
mercury at an‘absolute temperature of 508°. The ratios corresponding
to other absolute temperatures of mercury were calculated from the
equation R’ = R + c(t’ — 2), the value of R being 1°6331, as given
above. Of course other ratios might have been read from the straight
line; but it seemed better to employ the factor 0°0004788 in the
calculation.

The data employed and the results of this calculation are as

follows :—

Temperature Absolute Absolute
(centi e). temperature temperature
grad of mercury. of water.

408° 1°585220 257 *38°
413 1°587614 260°14
418 1 °590008 262°89
423 1°592402 265° 64
428 1°594796 268°37
433 1°597190 271°
438 1°599584 273° 8:3
443 1°601978 276°
448 1°604372 279°
453 1 606766 281°
458 1°609160 284°
463 1°611554 287°
468 1°613948 289 °¢
473 1°616342 292°
478 1°618736 295°
483 1°621130 297°
1°623524 300°

1 °625918 303°
1°628312 305°
1°630706 308 °
1°633100 311°
1°635494 313°6
1°637888 316°
1°640282 318° 74592
1°642676 321° 85 °010
1 *645070 324°00 96 *661

VAPOUR-PRESSURES OF MERCURY.

Absolute Absolute
—— temperature temperature
(centig e). of mercury. of water.

Vapour-
pressure.

265° 538° 1°647464 326°55° 109°556 mm.
270 *649858 329°11 123 *905
275 652252 331° 66 139-802
280 654646 334°21 157 *378
285 *657040 336°75 176 °733
290 "659434 339°28 197°982
295 *661828 341 *80 221°251
300 "664222 344°31 246°704
305 666616 346°81 274 443
310 *669010 349°31 304.°794
315 671404 351 °80 337 °753
320 *673798 354 °28 373 528
325 *676192 356 °75 412°249
330 678586 359° 22 454 °277
335 *680980 361°68 499 656
340 683374 364°14 548 °715
345 685768 366 *59 601°583
350 688162 369° 04 658°515
355 *690556 371°48 719°772
360 *692950 373°91 785 *107
365 695344 376° 33 855° 223
697738 378 °75 930°335
*700132 381°16 1010°47
*702526 383 *56 1096° 22
*704920 385 °95 1186 °67
*707314 388°34 1283°71
*709708 390°72 1386°60
*712102 393°09 1495°60
*714496 395 °45 1611°19
*716890 397°81 1733 -79
*719284 400°16 1863 °36
721678 402°50 2000°21
*724072 404° 85 2145°57
1°726466 407°19 2298 ° 80
1 °728860 409°51 2459°41
1°731254 411°83 2628°79
1°733648 414°15 2807 *53
*736042 416°47 2996 ° 06
418°77 3192°69
421°07 3399-50
423° 36 3616 °22
425° 65 3843 -68
427°92 4080°10
430°18 4327°14
432° 44 4585 * 95
*755194 434°70 4856 °74
*757588 436° 96 5139 °89
*759982 439 21 5434°99
*762376 441°45 5741°86
“764770 443 °67 6059 -16
767164 445-90 6391°49
*769558 448-12 6736-60

1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1

Pst fee et et et et feed fet feed fe fed et et

In calculating pressures below 46 mm., Regnault’s formula

48 RAMSAY AND YOUNG ON THE

expressing the relation between temperature and pressure of water
between 0° and 100° has been employed.

As it was impossible to calculate the vapour-pressure of mercury
below 135°, owing to our ignorance of the vapour-pressure of water
at low temperatures; and as a knowledge of the lower vapour-
pressures of mercury is necessary, the constants of a formula of the
form employed by Regnault in calculating the vapour-pressures of
water below 0° were calculated. The formula is—

log p = a + bat.

In calculating this formula, the temperatures and pressures made
use of in determining the value of the constants were as follows :—

Temperatures. Pressures.

160° 4013 mm.
220 31°957_ ,,
280 157°378 ,,

The constants are therefore—

a = 4493745 log b = 0°5899797
b = — 3890276.
t =t°C.—160 log a = 19980929.

The table which follows shows the vapour-pressures of mercury at
temperatures below 160°, calculated as above described.

Temp. | Pressure. | " Temp. Pressure.

40° 0°008 mm. oe " 140° 1°763 mm.
50 in , 145 2°181 ,,
60 ‘ “71S 150 2°684 _,,
70 "05 . . 155 3-289 _,,
80 “092 “416 160 4°013 ,,
90 "16 .

As these numbers are extrapolated from 220°, it appeared neces-
sary to control them by a determination at a lower temperature. The
vapour-pressure of mercury was therefore directly determined at the
boiling point of chlorobenzene under a pressure of 754°2 mm., corre-
sponding to a temperature of 131°8°, by the method employed for its
determination at the boiling point of methyl salicylate. Twenty-
three readings were taken, the level of the mercury in the vapour-
pressure tube being altered from time to time. The mean result was
1°58 mm.; the mean probable error was 0°14 mm. The calculated
pressure is 1:24 mm.; but we think that the experimental result is as

— @

ona

“e+e

VAPOUR-PRESSUKES OF MERCURY.

close as can be expected from the very small amount to be measured,
and we regard it as a sufficient confirmation of the accuracy of the
table given above. At 132° Regnault’s pressure is 2°30 mm., and this
is manifestly too high.

In order to afford data whereby the temperatures of mercury
vapour used as a jacket may be easily ascertained from the read
pressure, we append a table of the vapour-pressures for each degree
centigrade. This table is to be substituted for the one published in
the Trans., 1885, p. 656.

The results given were smoothed by the method of differences, and
each degree was calculated by differences.

Temp. Pressure. Temp. Pressure. Temp. Pressure.
270° 123 ‘92 mm. 801° 252-18 mm. 332° 472-12 mm.
271 126°97 ,, 302 257°65 ,, 333 481°19 ,,
272 130-08 ,, 303 263-21 ,, 334 490 40 ,,
273 ~—s|:133 26 ,, 304 268 ‘87 ,, 335 499°74 ,,
274 | 136°50 ,, 305 274°63_,, 336 509°22 ,,
275 | 139-81 ,, 306 280748 ,, 337 51885 ,,
276 ~=—s|:143°18_,, 307 | 286°43 __,, 338 528°63_,,
277 | 146-61 ,, 308 | 292°49 ,, 339 538°56 ,,
278 ~=—s |:150°12._,, 309 «| 298-66 ,, 340 548-64 ,,
279 | 153°70_,, 310 304°93 ,, 341 558°87_,,
280 =| :157°35 _,, 311 311-30 ,, 342 569°25 ,,
281 + =+| 161-07 ,, 312 =| 317-78 ,, 343 579-78 _,,
282 | 164°86 ,, 313 324°37 ,, 344 590-48,
283 | 168°73_,, 314 331-08 ,, 345 601°33_,,
284 172°67 ,, 315 337-89 ,, 346 612°34 ,,
285 | 176°79 ,, 316 344°81 _,, 347 623°51_,,
286 =| 180°88_,, 317 351°85_,, 348 634°85_,,
287 | 185°05 _,, 318 | 359-00 ,, 349 646°36_,,
288 ~—s||:189°30 ,, 3i9 366 °28 _,, 350 658°03_,,
289 | 193°63__,, 320 | 878°67 ,, 351 669°86__,,
290 | 198-04 ,, 321 381°18 _,, 352 681-86 ,,
291 202-53, 322 388°8l ,, 353 694°04 ,,
2902 =|: 207°10 ~,, 323 396°56__,, 354 706°40 ,,
293 =|: 211-76 ,, 324 404°43 ,, 355 718°94 ,,
294 216°50 ,, 325 412-44 ,, 356 731-65 _,,
295 | 221-33 ,, 26 420°58 _,, 357 744°54 ,,
296 | 226°25 ,, 327 428 83 ,, 358 757°61 ,,
297 = |: 231°25—,, 328 437-22 ,, 359 770 ‘87 ,,
298 =|: 236-34 ,, 329 415°75 ,, 360 784°31 ,,
299 =| 241-53, 330 454°41 ,,

300 246°81 ,, 331 463°20 ,,

The result of this investigation, we venture to think, furnishes a
most convincing proof of the justice of the generalisations on which
it is based.

VOL. XIIX.

50 JAMES: ACTION OF PHOSPHORUS PENTACHLORIDE

ADDENDUM.

Since reading the above paper, Professor Herbert McLeod has
kindly directed our attention to two memoirs on the same subject, one
by Hagen (Ann. Phys. Chem., N.F., 16, 610), and one by Hertz (ibid.,
N.F., 17,193). The methods employed by both these experimenters in
measuring temperature are open to criticism. Hagen’s experiments
appear to have been conducted with very great care; but the results
are abnormal, for he finds a pressure which must be greatly in excess
of the truth at low temperatures. Hertz employed a method analo-
gous to ours, and the closeness of the coincidence between his results
and ours is remarkable, when it is remembered that ours are extra-
polated from measurements at much higher temperatures, the correct-
ness of which was confirmed by their agreement with the relations
mentioned in our paper. We think it advisable to reproduce their
numbers at low temperatures, comparing them with those of Regnault
and ourselves.

Temperature. Regnault. Hagen. |Ramsay and Young.

0-015 ‘ 9 |
0-018

saseseges! ||
COoOocnwnonre

0:
0°02
0:
0°
0:
0°
0:3:
-| Orf
.| 0°27:
Bt
«| O
| 5°

—
S2OPRrROOCSOCOOoOCoCSO
FIO PRROSOOSOOCOSCSO

ore
oouos
Om ce
io

VIL.—Action of Phosphorus Pentachloride on Ethylic Diethylaceto-
acetate.

By J. Wituiam James, University College of South Wales, Cardiff.

THe action of phosphorus pentachloride on ethylic acetoacetate was
first carefully investigated by Geuther (Jenaische Zeitschrift, 4,
Heft 4), who found that the ethoxyl-group was displaced by chlorine,
giving rise to the chlorides of two isomeric chlorocrotonic acids.

ON ETHYLIC DIETHYLACETOACETATE. 51

Some years later, Riicker (Annalen, 201, 56) submitted an ethereal
monalkylacetoacetate to the same treatment, and again found that
the ethoxyl of the carboxyl-group was displaced by chlorine, the
product being the chloride of a single monalkylchlorocrotonic acid.
Judging from these results, it was to be expected that the action of
phosphorus pentachloride on a dialkylacetoacetate would give rise to
the formation of the chloride of a dialkylchlorcrotonic acid ; this,
however, is not the case, the chief result of the reaction being a
mono-, and a di-chloro-substitution product, together with a small
quantity of a monalkylchlorcrotonic ethylic salt, as the following
experimental results testify.

100 grams (1 mol.) of ethylic diethylacetoacetate boiling at 205—
215° was poured into a flask containing 225 grams of phosphorus
pentachloride (2 mols.). No reaction took place at the ordinary tem-
perature or at 100°; the mixture was therefore gently heated in con-
nection with a reflux condenser until all the chloride had dissolved,
and was kept boiling gently for some time. The upper end of the con-
denser contained a calcium chloride tube, which entered a cylinder filled
half full of water. A thermometer placed in the vapour of the boiling
liquid remained very constant at 85°, so the temperature in the liquid
could hardly have been much over 100°. During the reaction, volumes
of hydrogen chloride were evolved, and also some ethyl chloride
which could be inflamed in the cylinder. After boiling for five hours,
the liquid had become pale yellow, and on cooling about 25 grams of
unused phosphorus pentachloride crystallised out. The liquid poured
off from the crystals was now distilled until the thermometer showed
120°, in order to remove most of the phosphorus oxychloride and
trichloride, of which it mainly consisted. The residue, which was
now brown in colour, was allowed to cool and poured into cold
water; and then, as it could not be boiled under ordinary pressure
without undergoing considerable decomposition, it was distilled in a
current of steam. The oil which passed over very slowly was heavier
than water, and insoluble in it. It was collected in three principal
fractions of equal volume (I, II, III). The first drops which distilled
possessed a peculiar odour, resembling that of the camphor compounds ;
this was collected separately, but owing to the small quantity it has
not been more closely examined; it boils below 200°, with slight
decomposition. The last portion was also collected apart (about
3 c.c., Fraction IV). Each of the above fractions was separated
from the water, dried over sulphuric acid, and afterwards analysed.

The aqueous distillate had a strongly acid reaction, but after
neutralising with sodium carbonate, evaporating to dryness and
exhausting with absolute alcohol, only a small quantity was dissolved,
the residue consisting of sodium chloride. The liquid in the flask

E 2

52 JAMES: ACTION OF PHOSPHORUS PENTACHLORIDE

also had a strongly acid reaction due to phosphoric and phosphorous
acids.

Fraction I.—A portion (a) of this slightly yellowish liquid, on dis-
tillation, boiled from 210° to 220° with decomposition, hydrogen chlo-
vide being ‘given off whilst a dark-brown residue remained in the
retort. This distillate was analysed, as was also another portion
which had not been distilled (b).

(a). 0°2685 gram substance produced 0°550 gram CO, and 0 2025

gram H.O.
02195 gram heated with CaO gave 0°118 gram AgCl.
(6). 0°255 gram substance gave 0°517 gram CO, and 0:187 gram H,0.
03145 gram finally produced 0°195 gram AgCl.

Calculated for
CH,Cl-CO-C(C2H;).-COOC.H,,.
ts ial

16°09

220°5 160-00

The substance, according to these analyses, consisted mainly of
ethylie diethylmonochloracetoacetate slightly contaminated with a
compound richer in carbon and hydrogen, probably free from
chlorine.

Ethylie diethylchloracetoacetate is a colourless liquid, insoluble in
water, and of a pleasant odour. It has a sp. gr. of 1°063 at 15°. It
is miscible in all proportions with alcohol, ether, and benzene.

Fraction I1.—This might consist of a mixture of the two substances
which composed fractions I and II], that is, a mixture of mono- and
di-chlorinated ethylic diethylacetoacetate, but also of ethylic ethyl-
chlorcrotonate, CH;-CCl : C(C,H;)*COOC,H;.

Two chlorine estimations, in two samples separately prepared, gave
21°8 and 211 per cent. Cl, which agree pretty well with the amount
of chlorine in this last-named ether, but also equally consistently for
a mixture of equal numbers of molecules of mono- and di-chlorinated
ethylic diethylacetoacetate.

This fraction was consequently again distilled in a current of steam
until one-half had passed over, dried over sulphuric acid, and the
chlorine determined.

Fraction Ila.

0288 gram substance produced 0°200 gram AgCl = 17:1 per
cent. Cl.

ON ETHYLIC DIETHYLACETOACETATE. 53

Fraction IIb.
0277 gram substance produced 0°265 gram AgCl = 23°6 per
cent. Cl.
Calculated for ethylic diethylmonochloracetoacetate 16°1 per cent.
Calculated for ethylic diethyldichloracetoacetate 27°8 per cent.
Presuming that fractional distillation in a current of steam can be
carried out with a fair amount of exactitude,* these results tend to
show that a compound was present which was neither the monochloro-
nor dichloro-substitution product, since in the first case the chlorine
found would be too high, and in the second too low. I have sub-
stantiated this supposition by again distilling Fraction IIb to one-
half, and analysing it.
Fraction IIb 1.
0'3245 gram substance produced 0°266 gram AgCl = 20°2 per

cent. Cl.
0'3140 gram substance produced 0°€200 gram CO, and 0°2005 gram

Calculated for Found.
CH;-CCl : C(C,H;)-COOC,H;. -——ot TF

a 1s z II.
44 — 53°85
7°36 7°09

20°11

Although the percentage of chlorine found agrees with the above
formula, still the carbon, and especially the hydrogen, are too low.
It is, I think, extremely probable that this substance consists mainly
of pure ethylic ethylchlorcrotonate, as the evolution of ethyl chloride
and the occurrence of phosphorus oxychloride can then be satisfac-
torily accounted for.

Fraction II1I.—Analyses have shown this fraction to be a dichloro-
substitution product of ethylic diethylacetoacetate.

I. 0°274 gram substance produced 0°297 gram AgCl.
IT. 0°3135 _,, “ - 05455 gram CO, and 0182
gram H,O.
The oil was now distilled to one-half, and that remaining in the

flask dried and analysed for chlorine. The percentage was found to
be nearly the same.

* In these operations a long-necked flask was used, and the tube in connection
with the condenser was about } in. in diameter.

54 JAMES: ACTION OF PHOSPHORUS PENTACHLORIDE

III. 0°2775 gram substance gave 0°319 gram AgCl.

Calculated for Found.
CHCI,-CO:0(C,H;).-COOC,H;. ‘ ~Anme
L al . II.

47°05 47°4
6°27 6°4

27°84

Ethylic diethyldichloracetoacetate is an oily, slightly yellow liquid,
having a pleasant odour. It is insoluble in water, but miscible with
alcohol and ether, and has at 15° a sp. gr. of 1:155. It cannot be dis-
tilled under the ordinary pressure.

Fraction 1V.—This liquid, consisting of the last drops which were
distilled, appears to contain about 50 per cent. of a trichloro-substi-
tution product of ethylic diethylacetoacetate, the percentage of
chlorine in which is 36°8.

0°2115 gram produced 0°2655 gram AgCl = 31-06 per cent. Cl.

Il. Formation of Oxyketones—Action of Sodium Methylate on Ethylic
Diethylchloracetoacetate.

A solution of 4:2 grams of sodium (1 mol.) in 40 grams of methyl
alcohol was poured into 40 grams of ethylic diethylmonochloraceto-
ucetate* contained in a flask; a reaction took place at once with
considerable elevation of temperature, the liquid becoming brown.
It was necessary to place the flask in cold water for a time, but after-
wards the mixture was heated on the water-bath, and finally in sealed
tubes for three hours at 100°. The contents of the tubes were then
distilled to dryness on an oil-bath, and the distillate fractionated.
Most of the methyl alcohol was thus separated, and the residue, which
boiled from about 80° upwards, was shaken with a saturated solution
of calcium chloride several times, and the upper layer separated and
submitted to fractional distillation. From this liquid, I succeeded in
obtaining two compounds ; one boiled at 150—132° and represented
the principal part, the other passed over between 185—190° and.
was obtained only in small quantity.

An analysis of the liquid boiling at 185—190°, which still contained
a little chlorine, gave the following numbers :—

* Obtained from 100 grams of diethylacetoacetate and PCl;. After distilling 1 or
2.c. in a current of steam, the remainder was collected until about two-fifths of
tue whole had passed over.

ON ETHYLIC DIETHYLACETOACETATE. 55

(2265 gram substance produced 0°5135 gram CO, and 0°1855 gram
H,0.
Calculated for
Cy, H 994.
ET.
132s 6111
20 921
64 29°68

216 100°00

This result agrees in the main with the composition of ethylic
methoxydiethylacetoacetate, CH,(OCH;)*CO-C(C,H;).,COOC,H;. It
is a pleasant smelling, colourless liquid, soluble in alcohol and ether,
and heavier than water, in which it is insoluble.

I have not been able to obtain this compound free from chlorine.
If an excess of sodium methylate is taken, in order to remove the
whole of the chlorine, the substance boiling from 130—132° is chiefly
produced.

Analysis of the body boiling from 130 —132° :—

I. 0°276 gram substance gave 0°6475 gram CO, and 0°268 gram

H,0.

II. 0°2745 gram substance, obtained from a second preparation,
gave 0'6465 gram CO, and 0°2725 gram H,0.

III. 0°2465 gram of the same substance gave 0°5820 gram CO, and
02415 gram H,0.

IV. 0320 gram of the portion boiling from 125—130°, which was
very little, gave 0°'7600 gram CO, and 0317 gram H,0.

Caleulated for Found.
C7H\ Ors. _ ~ —,
nooo TF k. IL. III. IV.
84 64°6 640 642 644 647
14 10°8 10°78 11°03 1084 11°0

32 24°6 = — — —_

130 = 100°0
These numbers agree pretty well with each other, and with the
formula of an oxyketone, viz., methoxymethyl-ethyl-acetone (methowy-
methyl butyl ketone), CH,(OCH;)-CO-CH(CH;)(C,H;). It might
certainly have been expected that a methoxy-diethyl-acetone would
have been formed, according to the following equations :—

56 JAMES: ACTION OF PHOSPHORUS PENTACHLORIDE

CH,Cl-CO-CEt,,COOEt + CH;ONa =
CH,(OCH,):CO-CEt,,COOEt + NaCl.

CH,OCH,)-CO-CEt,COOEt + H,O =
CH,(OCH,)-CO-CHEt, + CO, + EtOH.

But this, according to the analytical results, does not appear to be the
case. The place of one ethyl-group has been taken by methyl, and
this displacement is probably brought about by the excess of methyl
alcohol. The researches of Geuther and Bachmann (Annalen, 218,
49) have already shown that in the case of the ketals (and this oxy-
ketone is possibly nearly related to them) a positive organic radicle
can be displaced by another poorer in carbon, e.g., by the action of
methyl alcohol on diethylacetal, methylethylacetal and dimethyl-
acetal are formed.

Methoxymethyl-ethyl-acetone is a mobile, colourless liquid, of sp. gr.
0°855 at 20°, very much resembling acetal. It has a burning taste
and a very pleasant odour, and it can be mixed with alcohol and
ether, but is insoluble in water. It boils at 130—132°, and the vapour
burns with a luminous flame tinged with blue. No compound with
sodium hydrogen sulphite could be obtained.

Action of Sodium Methylate on Ethylic Diethyldichloracetoacetate.

A solution of 9°5 grams of sodium (2 mols.) in 70 grams of methyl
alcohol was poured into a flask containing 50 grams of pure ethylic
diethyldichloracetoacetate, and the mixture treated precisely as before
described. In this case also, I finally obtained two liquids boiling at
134—135° and 190—200°; this last portion was very little, and
consisted chiefly of ethylic dimethoxydiethylacetoacetate, as the fol-
lowing analysis shows :—

0°211 gram substance gave 0°4445 gram CO, and 0°165 gram H,0O.

Calculated for
—_—_—_ Found.

2
‘144-5853 57°4
22 87
80 325: oa

246 100°00

1 have endeavoured, in vain, to obtain this substance in a purer
state; it always contains a little chlorine, and since it is slightly
decomposed by distillation, the chlorine compound cannot be got rid
of in this way. An excess of sodium methylate causes a further
decomposition, as in the case of the monochlorinated derivative.

ON ETHYLIC DIETHYLACETOACETATE,

Ethylic dimethoxydiethylacetoacetate,
CH(OCH;),"CO-CEt.,COO Et,

is a pleasant smelling, colourless liquid, heavier than water, in which
it is insoluble ; it boils with partial decomposition about 195°
An analysis of the substance boiling at 134—135° gave the fol-
lowing result :—
I. 0°3235 gram substance gave 0°7335 gram CO, and 0°3065 gram
H,0.
II. 0°239 gram substance gave 0°5455 gram CO, and 0°225 gram
H,0.
Calculated for , Found.

"eats sage
T Ee

6183 62-24
10°52 10-46

These numbers correspond very closely with the above formula,
which represents a dimethoxry-diethyl-acetone,

CH(OCH;).*CO-CH(C,Hs).
The following equations explain its formation :—

CHCl,CO-CEt,,COOEt + 2CH;ONa =
CH(OCH;),CO-CEt,-COO Et.

CH(OCH;).CO-CEt,,COOEt + H,O =
CH(OCH,),,CO-CHEt, + CO, + EtOH.

It is a colourless, mobile, pleasant smelling liquid, having a burning
taste. It boils at 134°, and has a sp. gr. of 0°836 at 15°. It is inso-
luble in water, but miscible with alcohol and ether in all proportions.
Its vapour burns with a luminous flame. No compounds are formed
with acetic anhydride or sodium hydrogen sulphite.

The yield of these two methoxyketones is by no means a good one;
this is chiefly because a considerable quantity passes over with the
methyl alcohol, and cannot be separated by fractional distiliation ;
still, by adding powdered calcium chloride to the methyl alcohol dis-
tillate until a syrup is formed, shaking with ether—which dissolves the
oxyketones—and then removing the ether on a water-bath at 50°, I
have succeeded in recovering a fair amount of these compounds.

I am not aware that a methoxyketone has been previously de-
scribed, although L. Henry has recently (Ber., 14, 2272) prepared the

58 O’SULLIVAN ON THE SUGARS OF SOME CEREALS

ethyl salt of the alcohol of pyroracemic acid, CH,-CO-CH,°OC.H;,,
which he obtained from propargyl ether by means of mercuric bromide
and water.

Ammonia and Ethylic Diethylacetoacetate.

By the action of concentrated aqueous ammonia on this substance,
it was expected that amides would be obtained analogous to those
produced from ethylic acetoacetate, at the same time its abnormal
reaction with phosphorus pentachloride seemed to predict failure, and
this prediction has been verified by experiment. No reaction takes
place at 120—130° in sealed tubes, and on heating still higher, finally
to 190—200°, the ethylic diethylacetoacetate becomes decomposed,
forming ammonium carbonate and diethylacetoue boiling at 135—137°.

VIIT.—On the Sugars of Some Cereals and of Germinated Grain.
By C. O’Suttivay, F.R.S.

Ix the analyses of the cereals, and of germinated grain hitherto
published, if sugar is mentioned as one of the constituents, the
particular sugar or sugars have not been specified... Kihneman,
Ber., 8, 202, indeed, states that he isolated 0°6 to 1 per cent. sucrose
(cane-sugar) from germinated barley ; the evidence for the statement,.
us published, is, however, unsatisfactory and unconvincing. As the
question of the presence or absence of a sugar or sugars, and a
determination of the particular variety thereof, is of considerable
practical importance, I may be permitted to record the results of an
investigation undertaken with this object in view.

I shall describe the experiments with barley, as the operations
employed in dealing with it answer also for the other cereals, raw and
germinated.

The ground grain, when treated with water, yields the sugars to that
solvent, but, at the same time, other compounds are dissolved, which
render it impossible to bring the solution into such a condition as
would admit of its being satisfactorily worked with. The amount of
material which it is necessary to take requires the employment of
much solvent, so that the washings are very dilute, and the solution
cannot be concentrated, unless, indeed, under diminished pressure,
without the development of much colour, and the formation of
decomposition products. I have, on another occasion, pointed out that

AND OF GERMINATED GRAIN. 59

alcohol, sp. gr. 0°90, yields a clean, fairly workable solution, but the
amount of mucidin and fibrin dissolved is so great that it is almost
impossible to concentrate the extract obtained with it by distillation,
because, when the greater part of the alcohol has passed over, the liquid ~
in the distilling flask froths up and is carried over in quantity. Strong
spirit, in which the albuminoids are insoluble, cannot be used with
advantage, as the sugars are not very soluble in it. In consequence
of these considerations, I found it necessary to proceed as follows.
2U0 grams finely ground barley (this was a convenient quantity
to work with), were introduced into a flask capable of holding
15 litres; 200 c.c. of alcohol (sp. gr. 0°9) added, and the mixture
allowed to stand for 24 hours. 400 c.c. alcohol, sp. gr. 0°84, were
then added, and the whole digested at 40° for four hours. The
solution was filtered while still warm, and the filtrate kept by itself
to cool; the residue being washed by decantation several times with
alcohol, sp. gr. 0°85—0°86. It is well to keep the first 200 or 300 c.c.
of washings by themselves, because they, as well as the first filtrate,
deposit a considerable quantity of albuminoid on cooling. When the
extract and the first portion of the washings had become clear, after
cooling, they were mixed with the final washings and submitted to
distillation, 300 or 400 c.c. at a time. Each portion was concentrated
to about 50 c.c., and the residues were collected and allowed to cool.
Much albuminoid separated, and after 24 hours the supernatant liquid
was bright. This was again decanted into a distilling flask, the insoluble
deposit washed with a little water, the washings added to the contents
of the flask, and the whole submitted to distillation until the residue
was less than 100 c.c. and free from alcohol. This was shaken
with a little aluminium hydroxide and filtered; the filtrate with the
washings was made up to 100 cc. As the extraction and the
washing had been carefully done, this 100 c.c. contained all the
ready formed sugars of the 200 grams barley: the solution was
sufficiently bright for optical observations. The sp. gr.* of the
solution was found to be 1°02445; it had an optical activity = 10-1
divisious,¢ in a 200 mm. tube of a Soleil-Scheibler saccharimeter, and
12°159 grams of it reduced 0°1943 gram CuO. 80 c.c. of the solution

* In speaking of specific gravity it will be understood that I mean the weight of
the solution compared with the weight of water, the bulks and temperatures being
equal, and the weight expressed in the ratio of water = 1. I make the comparison
15 °5°

= 1.
15°5

t+ In this paper, I employ the divisions of a Soleil-Scheibler instrument to
express the optical activity, because they answer the purpose as well as angular
values, and the trouble of calculation is thereby avoided. The observations are all

-.-0

made in a 200 mm. tube and at a temperature of 15°3°.

at 15°5°, and this may be indicated sp gr.

60 O’SULLIVAN ON THE SUGARS OF SOME CEREALS

were digested in a flask with 0°03 gram invertase for 4 hours at
52—54° ;* after cooling, the contents of the flask were made up to the
original bulk (80 c.c.) by the addition of a few drops of water, and
filtered. The solution was then sufficiently bright for an optical
observation. Occasionally, however, the filtrate, obtained in the way
indicated, from some material, requires to be treated with a little
animal charcoal ; in such cases, the least possible quantity should be
used. The sp. gr. of the filtered solution was found to have
increased to 1°02525, the optical activity to have diminished to
— 251 divs., and the K to have increased—3:146 grams solution
reduced 0'243 gram CuO.

From these data, the K of the original solution was = 0°74 gram
dextrose, which was increased to 3°58 grams by the action of
invertase ;. the loss in optical activity on inversion was represented by
12°61 divs. (10°1 + 2°51); and the increase of solid matter, as indicated
by the sp. gr., was = 0°21 gram.t The original solution, therefore, con-
tained a substance capable of being acted upon by invertase, yielding
thereby a substance of increased reducing power, of diminished
optical activity, and of increased sp. gr. These are some of the
properties of sucrose; in fact, there is no other known sugar acted
upon by invertase in the way described. Let us see how the factors
observed agree with those belonging to that sugar—how the increase
in reducing power agrees with the diminution in optical activity.
The 2:84 grams (3:58 — 0°74) of reduction would be yielded by the
inversion of 2°69 grams sucrose, because

105°25 : 100°: : 2°84 : 2°69,

the first term being the amount of invert sugar yielded by 100 grams
of cane-sugar. Now | gram sucrose in 100 c.c. solution, when acted
upon by invertase, loses an optical activity = 5°18 divs., the observed
loss was 12°61 divs.; hence, 12°61 + 5°18 = 2°43, the grams of
sucrose in the 100 c.c. solution. We thus see that—

the optical activity indicates 2°43 grams sucrose, and
the reduction - 269 ,, sucrose.

Further, 2°69 grams sucrose, on being inverted, take up 0°15 gram
water; the observed increase in sp. gr. indicates 0°21 gram, including
the invertase. In this experiment the agreement between the sucrose,
calculated from the optical activity and the reduction, is not as close
as usual; in another experiment

the optical activity yielded 2°79 grams sucrose, and

the reduction ‘ 2°77 +, +=sucrose ;

* If sulphuric acid were employed, it would act not only upon the sucrose, but
also upon maltose and dextrose, should they be present.
+ Part of this is, no doubt, due to concentration during filtra tion.

AND OF GERMINATED GRAIN, 61

and the results of very many determinations lie between these two
extremes. Hence the evidence for the presence of sucrose is as con-
clusive as need be desired.

Having obtained this evidence, we may turn to an examination
of the reducing power in the original 100 c.c. solution, and see if we
can determine to what sugar or sugars it was due.

It may have been due to inverted sucrose ; if so, the optical activity of
the amount of reduction, taken as invert sugar, plus the optical activity
of the sucrose found, should be equal to the optical activity observed.
The original reduction corresponds to 0°74 gram invert sugar, and we
may take the sucrose at 2°69 grams; now 1 gram sucrose in 100 c.c.
solution is equal to 3°84 divs., and 1 gram invert sugar to — 1:23 divs. ;
hence 2°69 x 3°84 = 10°33 less 0°74 x 1:23 = 0°91 = 9°42, the number
of divisions corresponding to such a mixture; the activity observed
was 10°1 divs. These numbers do not agree sufficiently closely to give
any support to the supposition, and, even if the approximation were
closer, the evidence for the inference would be far from con-
clusive. Of course, if the quantity of sucrose indicated by the optical
activity were taken for the purpose of the calculation, the difference
between the observed and calculated numbers would be greater.

A mixture of levulose and dextrose could be worked out to fit in
with the observed optical activity, but this would be in no way satis-
factory; I, therefore, tried by other means to obtain evidence of the
character of the sugar.

66:939 grams of the inverted solution were sterilised by boiling and
cooled ; 0°4 gram pressed yeast was added to them, and fermentation
allowed to take place, first at the ordinary temperature, and then at
20—22°. The apparatus in which this operation was conducted con-
sisted of a flask capable of holding 120 c.c., in which was the fer-
menting liquid, and two small wash-bottles, through a few c.c. of
water in each of which the carbon dioxide evolved was made to pass.
At the end of 10 days, when all action had ceased, the fermented
liquid was transferred to a distilling flask, the first rinsings being
made with the water of the wash-bottles, and further -with a few c.c.
water, care being taken that the liquid and washings did not much
exceed 100 c.c. The distillate was collected in a 100 c.c. vessel, as is
done in the usual method of determining original gravities. When
nearly 100 c.c. had passed over the operation was stopped, and the
distillate made up at 15°5° to 100 cc. Of this the sp. gr. was found
to be 0°99772, which, according to Fownes’ tables, is equal to 1°22
grams absolute alcohol in the 100 c.c. Now as only 66°959 grams
original solution, sp. gr. 1:02525, were taken, we arrive at the quantity
of alcohol derivable from the whole 100 c.c. by the proportion —

66°94 : 102°53 : : 1:22 grams : 1°87 grams;

Bana v4

APA 73

—
<—
ae
& +
e 4

-*
s® Bb

62 O’SULLIVAN ON THE SUGARS OF SOME CEREALS

that is, the original 100 c.c. solution would have yielded 1°87 grams
absolute alcohol. According to the best available determinations
(Pasteur, Ann. Chim. Phys., 58, 323), 100 grams of a C,H,,0, sugar
yield 48°5 grams alcohol; the 1°87 gram is, therefore, the product of
3°86 grams sugar, for—

48°5 : 100: : 1:87 : 3°86.

The reducing power in the inverted solution was only = 3°58 grams
of dextrose; hence, we have here an indication that a sugar with a
less reducing power than invert sugar, or one with no reducing power
at all, had fermented.

Let us now compare these numbers with the amount of solid matter
that disappeared from the solution during fermentation. The contents
of the distilling flask were transferred to a 100 c.c. measure, and
with the washings, were made up to the 100 c.c. mark at 15°5°.
The sp. gr. was found to be, with the yeast in, 100676, and, after the
yeast was separated by filtration, 1:00620. The 0-4 gram yeast
originally added contained 0°13 gram solid matter, whereas the
presence of the yeast in the fermented liquid was indicated by a
sp. gr. of 1:00056 = about 0°14 gram; consequently, the yeast may be
considered to have done its work without having taken anything

from, or yielding anything to the solution. Had the whole 100 c.c.
been taken, the sp. gr. of the residue would have been 1°00948,

because—

66°94 : 102°53 : : 620 : 948;

hence, 102525 — 0:00948 = 101577, the sp. gr. in 100 c.c. of the matter
fermented. If we take 1-00385 as representing the gravity of 1 gram
sugar in 100 ec.c. solution, the sp. gr. of the matter fermented
indicates 4°09 grams, and this without taking into account the
glycerin, succinic acid, &c., produced. We have then the following
numbers for the sugars in the inverted solution :—

3°58 grams from reduction,
», the alcohol on fermentation, and
409 ,, », the gravity lost on fermentation.

The numbers calculated from the alcohol and from the gravity lost
on fermentation should agree better than they do in this case, and I
may say, they usually do, but these results afford a clear indication of
the facts generally observed. Without giving preference to any one
number, we may, for our purpose, take that derived from the alcohol
as the amount of sugar that had fermented. When we compare this
with the amount of sugar represented by the reducing power, we
observe that at least 0°29 gram (3°87 — 3°58) of matter which had no

AND OF GERMINATED GRAIN. 63

reducing power fermented. This is a property of maltose, but the
body may be any non-reducing sugar not acted upon by invertase.

Before we enter, however, into a consideration of this part of the
subject, I must mention that the unfermented residue had no optical
activity, and that 25 c.c. of it, boiled with 10 c.c. Fehling’s solution
properly diluted, gave no trace of reduction even after 15 minutes’
boiling; a slight precipitate did form, it is true, but this was flocculent
and white ; hence, all the sugars had fermented. This result is not
always obtainable; the residue very frequently exhibits an optical
activity, and possesses a reducing power. Sometimes, the relation of
the reducing power to the optical activity is such as to indicate the
presence of maltose, but, as a rule, the relation holds good for no
known sugar, and the optical activity is high for the redaction,
higher than would be indicated by dextrose, but no definite factors
can be arrived at, the quantities are too small. The residual optical
activity being a + quantity, the sugar cannot be levulose, and it
must be less fermentable than that sugar, or it would not have been
left behind whilst the invert sugar disappeared.

Now, to return to the matter fermented: it seems at first sight as
if the 0°29 gram of non-reducing power, described above as disap-
pearing during fermentation, might be due to the presence and
fermentation of 0°77 gram maltose (37°5, the non-reducing power of
maltose : 0°29 : : 100: 0°77), but when we consider that the optical
activity due to the sucrose is + 10°33 divs., whilst the original activity
was only = 1071, we see that the sugars other than sucrose must
have a minus power = — 0°23, and, therefore, this could not under
any circumstances be due to maltose alone, and probably not to any
mixture of maltose with other sugars. If it were not for the disap-
pearance of this non-reducing matter during fermentation, it would
appear probable that the original reducing power, judging from the
optical activity, is due to dextrose and levulose, the former being in
eXcess.

I may state that when barley has been extracted with alcohol as
described, if the residue is treated with water, the aqueous solution
concentrated, and the amylans and other substances precipitated by
alcohol, sp. gr. 0°84, the alcoholic supernatant liquid contains no
optically active or reducing substance. This I proved by distilling off
the alcohol, and concentrating the residue to a small bulk. All the
sugars of barley can, therefore, be dissolved by careful treatment with
alcohol as described above.

Such are the general indications regarding the sugars in barley.
The results of the examination of at least 20 varieties may be sum-
marised as follows :—

1st. The diminution in optical activity and the increase in reducing

64 O'SULLIVAN ON THE SUGARS OF SOME CEREALS

power, produced by the action of invertase, agree well with the
factors of sucrose, and the numbers obtained show that barley
contains between 0°8 and 1°6 per cent. of that body.

2nd. There is a variable quantity of a sugar or sugars present
which reduce less than dextrose, and the optical activity of which is
always a minus quantity, but I have not been able to establish any
satisfactory constant relation between the optical activity and K.

3rd. Frequently the solution containing the sugars does not
completely ferment ; when this is the case the residue has a positive
optical activity and a K which are variable, and cannot be referred to
any known sugar. The substance is less fermentable than levalose.

Wheat, treated in the same way, yields not more than 0°5 per cent.
sucrose, but there are indications of a moderately high levorotary,
non-reducing, fermentable sugar of which I hope to be able to say
something at a future time.

In dealing with germinated barley (malt), I found 100 grams
sufficient for the purpose, but as it is well to have at least 125 c.c. of
solution, I employ 125 grams substance.

125 grams malt, finely ground, were digested at 40° with 500 c.c.
alcohol sp. gr. 0°9 for six hours; to this 700 c.c. of alcohol, sp. gr. 0°84,
were added, and the mixture allowed to stand for 24 hours. The

clear supernatant liquid was decanted through a filter, and the
residue washed gradually with 400 to 500 c.c. of alcohol of the last-
named strength. The filtrate was then treated in the same way as is
described in the case of barley, and 125 c.c. of fairly colourless, bright,
aqueous solution obtained. The sp. gr. of this solution was 1°4890,
and its optical activity = 35°5 divs. In determining the reduction—

Ist Exp. 2°793 grams solution gave 0 207 gram CuO.
2nd Exp. 27920 ___,, ~ » O2l” ,, CuO.

These numbers correspond to—

lst Exp. 3°52 grams dextrose in 100 c.c. solution, and
2nd Exp. 3°53 __s,, ~ - solution.

To 100 c.c. of the solution, 0°06 gram invertase was added, and the
whole digested at 50—53° for four hours. This was sufficient to com-
pletely invert any sucrose that may have been present. The solution
was cooled to 15°5°, and then made up with a few drops of water to
100 c.c. The sp. gr. of this solution was 1:05020, and its optical
activity = 112 divs. :—

Ist Exp. 1°322 grams of it gave 0°239 gram CuO, and
2nd Exp. 1124 - » 0204 , Cu.

AND OF GERMINATED GRAIN.

These numbers correspond to—
lst Exp. 8°61 grams dextrose in 100 c.c. solution, and
2nd Exp. 8°64 __,, “ ‘i solution.

The increase in reduction is, therefore, equal to 5°11 (8°64 — 3°53
= 5°11) grams dextrose or invert sugar, corresponding to 4°85 grams
(ros38 = 485) sucrose.

1:0526

The optical activity in the original solution was 35°5 divs. ; after
inversion, it was 11-2 divs. ; there was consequently a loss of 24°3 divs. ;
a loss of 5°18 divs. corresponds to 1 gram sucrose in 100 e.c., and

243 = 4°69, the number of grams sucrose indicated by the decrease

5°
in optical activity.

We have then, in the 100 c.c. solution, 4°85 grams sucrose indicated
by the increase of reduction, and 4°69 grams sucrose indicated by the
decrease in optical activity.

These numbers, it will be observed, are percentages on the malt
taken.

In order to obtain some knowledge of the character of the reducing
bodies in the original solution, the inverted solution was submitted
to fermentation. Before examining the results of this experiment,
it may be as well to consider with what compounds we are likely to
have to deal.

We may take it that the original solution contained 47 grams
sucrose in 100 c.c.; the optical activity due to this is 18°05 divs.
(47 x 3°84); hence, 35°5 — 18°05 = 17°45 divs., are due to the sugars
represented by the 3°53 grams of reduction. If the reduction be taken
as dextrose, the optical activity is too high, if as maltose, too low; a
mixture of maltose and dextrose could be calculated to agree with it,
and one in which levulose was also present could be made to fit in,
but as we have no evidence of the presence of either sugar at present,
we must turn to the fermentation experiment to throw some light on
the subject.

102°4 grams of the inverted solution were sterilised and submitted
to fermentation with 0°5 gram pressed yeast, the usual precautions
being taken to make the carbon dioxide evolved pass through two
wash-bottles containing a little water. Towards the end, it was neces-
sary to add 0-002 to 0°003 gram active diastase to ensure the complete
fermentation of the optically active and reducing bodies.

The fermentation finished, the solution yielded on distillation—

100 c.c. alcohol, sp. gr. 0°99193, and
100 c.c. residue, sp. gr. 1°01473 with yeast in, and
sp. gr. 1:01392 ” out.
VOL. XLIX.

66 U’SULLIVAN ON THE SUGARS OF SOME CEREALS

The clear filtered residue was optically inactive and had no reducing
power.

The sp. gr. of the distillate indicates 4°50 grams alcohol in the
100 c.c., and if it be admitted that this is derived from a C,H,,0,
sugar, we get from the proportion

48'°5 : 100: : 45: 9°28
the number of grams of sugar whence it was derived. Now as only
102°4 grams of the original solution were employed, the proportion
102°4 : 105°02 : : 9°28 : 9°51

gives the grams fermentable in that solution. On the other hand
the total reduction was equal only to 8°64 grams dextrose; hence, it
would appear from this, that 0°87 gram of matter not indicated by
reducing power fermented, but as we have no evidence that all the
alcohol was derived from a C,H,,0, sugar, the factor is not of so
much value as is desirable. Let us see what can be made of the
difference between the original sp. gr. of the solution and that of the
residue. :

As pointed out above, the 102°4 grams solution gave 100 c.c. residue,
sp. gr. 1:01473 with yeast in, and 1°01392 after the yeast was sepa-
rated by filtration. The amount of yeast in the residue was repre-
sented, therefore, by 100 c.c., sp. gr. 100081. 4°264 grams of the
pressed yeast employed, suspended in 100 c.c. water, gave a sp. gr.
= 1:00466; the 0°5 gram used would, therefore, give a sp. gr. =
1:00054, thus leaving a quantity of matter represented by 100 c.c.
solution, sp. gr. 1°00027 (0°00081 — 0°00054) : this does not amount
to more than 0°07 gram, and as we do not know at present whence
it is derived, we may neglect it. The unfermented residue that was
left on the fermentation of the 1024 grams inverted solution
amounted to 100 c.c. of sp. gr. 101392, which, calculated out in the
same way as was the barley residue, gives the matter disappearing
during fermentation as 9°33 grams; but this figure can, as was
before shown, only be taken as an approximation. The number calca-
lated from the alcohol is 9°51, but this may be too high, a Cp.H»Oy,
sugar may have yielded a portion of it, whereas that derived from the
gravity must be too low, as the glycerin, succinic acid, &c., were not
allowed for. If, therefore, the mean of the numbers be taken, we
cannot be far from the truth ; in this way we arrive at 9°42 grams as
the amount of matter fermented.

The reduction before fermentation represented 8°64 grams sugar,
and as 9°42 grams fermented, 0°78 gram of matter without reducing
power must have disappeared. If this be attributed to maltose, we
have 2°08 grams of that compound, because

37°5 : 100: : 0°78 : 2°08;

AND OF GERMINATED GRAIN. 67

and 9°42 — 2°08 = 7°34 grams other sugars. 4°7 grams sucrose yield
494 grams invert sugar (reduction), and 2°08 grams maltose give
1:30 grams reduction; consequently, the original solution contained
2'40 grams of other reducing sugars—

8°64 — (4°94 + 1:30) = 2°40.

Admitting the quantity of sucrose and maltose to be correct, let us
see what optical activity will belong to these sugars.

Sucrose 47 x 3:84 = 18°05 divs.
Maltose 2°08 x 8°02 = 16°68 ,,

34°73,

The observed activity was 35°5 divs., thus leaving a dextro-power
= 0°77 divs. for the 2°4 grams sugar. These may be levulose and
dextrose, the latter being in excess. The quantity of each, in 100 c.c.,
may be calculated thus :—

: nae ; aS a = 1°65 gram dextrose,
and 2°4 — 1°65 = 0°75 gram levulose; — 5°52 and 2°96 being the
number of divisions due to 1 gram levulose and 1 gram dextrose
respectively in 100 c.c. solution; 2°4 the total sugar in grams, and
0'77 the optical activity in divisions due to them.

From these data, the malt employed contained—

4°70 per cent.

9°18 -
0°24 for the hydration of sucrose,
and we get 9°42 the amount of matter that
disappeared during fermentation.

Results similar to these have been obtained for all the malts worked
with.

I have tried to give, as clearly as possible, the evidence for the
presence of each sugar; that for the sucrose leaves little to be
desired, it is satisfactory ; with regard to the maltose the evidence is
not so convincing, for, although a certain quantity of matter without
reducing power disappears during fermentation, and the solution has
a sufficient optical activity to admit of the presence of the amount of
maltose calculated therefrom, as we have no means of checking one
result by the other, we cannot say that the observed facts are due

68 O’SULLIVAN ON THE SUGARS OF SOME CEREALS

absolutely to maltose; of that for the levulose and dextrose all that
can be said is, that there is a reducing power with which certain
mixtures of them correspond, and, that when the cane-sugar is low
in a malt, its place is taken by a reducing power and optical activity
which can be referred to levulose and dextrose.

I shall, however, show that it is not difficult to strengthen the
evidence for the presence of maltose. In nearly all cases, in the
fermentation experiments, unless a little active diastase had been added
to the solution, it was, after the fermentation had ceased, optically
active and reduced copper solution, the optical activity being, as a
rule, but not always, equal to that of maltose calculated from the
reduction. This is usually the case when the fermented solution
gives a reduction* equal to from 0°6 to 0°8 gram dextrose, correspond-
ing to 09 to 1:1 of maltose in. 100 c.c. Sometimes, indeed, the
optical activity is greater than the maltose calculated from the reduc-
tion ; this may be attributed to a little dextrin or malto-dextrin, and
it very probably is due to one or both, but I have no farther evidence
to prove it. I met with the indication only in a few malts.

Before summing up the results, I should state that it is very diffi-
cult to free malts from sugars by treatment with alcohol; indeed,
although I have made at least 20 estimations, I have only succeeded
in a few cases in obtaining a residue perfectly free from sugar. To
determine the point, the residue was treated with water at 40°, the
extract evaporated to a small bulk, which is, in this case, easily
accomplished without the development of much colour, and 40 to
50 c.c. hot alcohol (0°85) added; this throws down much flocculent
matter, and the clear supernatant liquid contains the sugar. This is
usually maltose, as invertase produces no change, and the reducing
power and optical activity corresponds to maltose. In this way, I have
found as much as 0°7 gram sugar left behind on treating 100 grams
malt. This may be due to the action of water and a little diastase
on the broken starch granules, but I think it is hardly so, as very
little, if any, of the transforming agent can be present.

The chief results obtained from the analyses of over 20 samples of
malt, from various sources, by the method described, may be sum-
marised as follows :—

* The estimation of the reduction in the fermented solution leaves much to be
desired, in consequence of the precipitation of other matter with the copper sub-
oxidé. When sufficient reduction is given by 12 to 15c.c. of the solution, and these
are heated to boiling, and added to the boiling dilute copper solution gradually,
the precipitate represents the reduction fairly accurately, but when the reducing
power is low, and 25 to 30 c.c. of the fermented solution have to be taken, the pre-
cipitate is decidedly impure; it blackens on drying, and, on ignition, burns like
tinder.

AND OF GERMINATED GRAIN.

Ist. Malts contain from—

2°8 to 6:0 per cent. sucrose,

1:3 to 50 me maltose,

1°5 to 30 “ dextrose, and
0°7 to 1°5 - levulose.

This does not mean that any variety contains only the maximum
or the minimum of all the sugars, but that the numbers given were
those observed in individual cases.

2nd. As a general rule, the sucrose and maltose are high in highly
germinated grain, and, when the sucrose is low in such grain, its
place is taken by the products of its inversion, levulose and
dextrose.

3rd. The numbers given as the results of the above described
experiment may be taken as fairly typical.

Without going into details, I may say that germinated wheat
yields similar results.

That some idea may be conveyed of the changes that take place in
the sugars during germination, I give the results of the analyses of
two barleys before and after germination, calculated per cent. on the
dry matter of the barley, the yield of dry malt being known.

Sugars. l
Before After
germination. | germination.

Before After
germination. | germination.

—_—_—— Ll

Sucrose........|
Maltose........ }
Dextrose ... a
Levulose .. j

From this, it is clear that during germination sucrose increases
largely, and that there is a decided production of maltose, dextrose,
and levulose. The high dextrose in No. 1 germinated barley is
frequently observed, but then, as in this case, the maltose is low, the
fact pointing apparently to a conversion of that body.

Starch disappears to a considerable extent during germination ;
from it the maltose is most probably derived. The source of the
sucrose, and of the products of its inversion, is not so easily indicated.

I need hardly point out that sucrose can be easily detected in
unfermented malt-wort by the loss of optical activity on addition of

* Sugars other than sucrose, by fermentation experiment.

O’SULLIVAN ON THE PRESENCE OF

invertase. This inversion also takes place immediately on the addition
of yeast, or in a short time if the wort be kept after exposure to the
air, doubtless from the growth of fortuitous yeast. In a wort
sp. gr. 1:065, a diminution of as few as 5 and as many as 10 divs.,
corresponding respectively to 5°7 and 11°4 per cent. sucrose on the
wort matter, have been observed on digestion at 50—52° with
invertase.

In conclusion, I may say that I have simply given the evidence
from which my inferences are drawn; this evidence must be taken
for what it appears to be worth; for myself, I shall not be satisfied
until each individual sugar is separated, crystallised, and examined by
itself. Further, I have as briefly as possible indicated how the data
were obtained ; if I were to give all the analytical numbers upon
which the inferences are based, I should have to occupy much more
space than the facts outside the inferences are worth.

My thanks are due to my brother James, and to my friend Mr. F.
W. Tompson, for much assistance in doing this work.

IX.—On the Presence of “ Raffinose” in Barley.

By C. O’Suttivan, F.R.S.

Ix the preceding communication, I pointed out that, although the
evidence for the presence of certain sugars in barley was moderately
conclusive, we could not be absolutely satisfied until each sugar was
isolated and examined by itself. So convinced was I of the truth of this
proposition that eight or ten years ago, early in my investigation on the
sugars of the cereals, I made an attempt to crystallise them. With
that object in view, 2 kilos. of ground barley were extracted with
alcohol in the same way as is described for the 200 grams in the last
paper. - The solution, freed from alcohol, was evaporated to a syrup,
which was dissolved in the least possible quantity of boiling alcohol,
sp. gr. 0°83. This solution, on cooling, deposited a syrup which was
again dissolved in just sufficient alcohol to hold it in solution when
cold; a little ether was then added to produce a slight turbidity, and
the whole put aside to crystallise. In a short time, cauliflower-like
segregations began to form, and after a few months they ceased to
increase. These were collected, washed with strong alcohol, and put
aside labelled “ Sucrose (cane-sugar) from barley,” for no other reason
than that they appeared under pretty much the same conditions as
those under which Ktihneman (Ber., 8, 202) said he isolated cane-

wTrFFens" wo 6S

“ RAFFINOSE” IN BARLEY. 71

sugar from malt. The quantity I obtained was not large; it did not
amount to more than 1°4 grams, or about 0°07 per cent. of the barley
employed. Although this was amply sufficient to determine whether
it was sucrose or not, so certain was I that it was, its crystal-
line appearance notwithstanding, that I did not think it necessary to
examine it.

Working with the barley of the season 1878, I believe, in the same
way, I obtained 4°55 grams of the cauliflower-like segregations from
5 kilos. material. This too I put aside, labelled ‘ Cane-sugar from
barley.”

Recently, while arranging the material for the preceding paper,
I thought it desirable to examine more closely the two preparations
described above. If they were sucrose, I had no doubt I could, by
recrystallisation, obtain recognisable crystals of that substance. Before
dissolving them, I examined a little of each preparation under the
microscope; I found the crystals were elongated, flattened prisms
terminated by a dome parallel to the shorter axis, the groups or
segregations consisting of the crystals radiating from a centre. Both
samples were alike, hence I had undoubtedly to do with a sub-
stance altogether different in crystalline form from sucrose. I pro-
ceeded to examine the preparations farther.

A determination of the optical activity of the dry matter in the one
gave [a]; = 125°, and in the other [a]; = 114°. Both preparations
were impure, for each contained a considerable quantity of ash, but
it was evident that I had the same substance to deal with.

I dissolved them together in a little water, filtered from some
insoluble matter, and added strong alcohol until a precipitate began
to appear. This solution, on standing, was filled with radiating
groups of beautiful silky crystals, which were collected, and washed
with alcohol so regulated in strength as to produce no turbidity in the
mother-liquid, the sugar being precipitated as a syrup on adding
strong alcohol to the concentrated solution.

The crystals are well-defined, flat, probably rhombic prisms,

Fie. 1. Fig. 2. Fie. 3.

72 O’SULLIVAN ON THE PRESENCE OF

terminated by a brachydome. Figs. 1, 2, and 3 give a fair idea
of their shape.

A second crop of crystals was obtained by concentrating the mother-
liquor to a syrup, and again adding a little strong alcohol.

On determining the optical activity of these, I found the results
did not quite agree; the first crop gave [a]; = 129°8° (c = 4142), and
the second [a]; = 132° (c = 2°392); both still contained ash. As,
however, the optical activity was so near, and the quantity of sub-
stance at my disposal so small, I mixed the two specimens, and sub-
mitted the mixture to recrystallisation.

A first and second crop of crystals were obtained as in the first
instance. I now found that both preparations had practically the
same optical activity, viz., for the substance dried in a vacuum
over sulphuric acid and then in dry air at 100° until the weight was
constant, [a]; = 134—135° (c = 4 to 5); hence, I concluded I had
purified the substance.

Ist Exp.—0-992 gram of the crystals, first allowed to remain in dry
air until the weight became constant, lost in a vacuum over sulphuric
acid 0°142, and then, in dry air at 100°, a farther 0-008 gram, making a
total loss of 0°150 gram.

2nd Exp.—2°061 grams, dried in the same way, lost a total of 0°312
gram.

These results indicate—

lst Exp. 15°12 per cent. water.
2nd Exp. 15°14 ™ =

Submitted to combustion in a stream of oxygen, 0°3148 gram dry
substance gave—
CO, = 0°4893 gram, and H,O = 01823;
0°0020 gram ash was left in the boat; this was chiefly potassium

carbonate.

The numbers give—
Theory for CyH,,Os.

42°66 per cent. 42°85 per cent.
6°47 ” 6°35 ”

These percentages are sufficient to indicate the empirical formula of
the body in the dry state, and, from the amount of water lost by the
crystals, we get the formula—

C,H;,.0,,23H.,0,
or by doubling it to eliminate the $ mol. H,0,
CisH206,5H,20,
which probably represents the molecular formula of the substance.

“ RAFFINOSE” IN BARLEY. 73

An aqueous solution containing 1 gram dry substance ‘in 100 c.c.
had a sp. gr. 1:003965; this shows conclusively that the substance as
burned did not contain any water of crystallisation.

The dry sugar absorbs water from the atmosphere and becomes a
glassy mass.

2°227 grams dry substance in 50 c.c. solution gave an optical
activity = 31°4 divisions of Soleil-Scheibler’s instrument; from this,
the specific rotary power for the dry sugar is [a]; = + 135°3°.

It does not reduce copper solution ; 0°1 gram boiled for 25 minutes
with 25 c.c. Fehling’s solution; properly diluted, yielded only a trace
of copper suboxide.

Invertase appears to act upon it, but very slowly: to 25 c.c. of a
solution possessing an optical activity =28:2 divs., 0°015 gram invertase
was added, and the mixture digested at 50—52° for three hours; the
optical activity of the cooled'solution was = 25 divs. A little more
invertase was added, and the digestion continued for another three
hours ; the activity had then fallen to 21 divs. This was a slow process,
but does not leave a doubt that invertase has an invertive action.
Had sucrose, in the same quantity, been in the solution, it would have
been all inverted in the first three hours. A one per cent. solution of
sulphuric acid at 100° reduced the optical activity, in one hour, from
28 divs. to 10 divs. In this case, the products of inversion had a
sp. rt. pr. [a]; = 43°5°, and'a K = 80°8. The amount of material at
my disposal did not warrant farther experiments in this direction, so
I did not continue then.

Treated with 4 parts nitric acid and 1 part water, it yielded a little
less than 30 per cent. on dry substance of an insoluble acid, which,
from its general behaviour and appearance ‘under the microscope, is
mucic acid. The filtrate from this acid, neutralised with ammonia
and acidified with acetic acid, yielded a precipitate of calcium oxalate
on the addition of calcium chloride. From these facts, it is highly
probable that galactose is one of the products of inversion, and, from
the composition, that two other sugars are produced. These must be
saccharic and oxalic acid yielding sugars.

The sugar is fermented by ordinary yeast.

I had worked with this sugar some’time under the name of
“cerealose,” but the recent papers of Scheibler (Ber., 18, 1779) and
Tollen (Ber., 18, 2611) leave no doubt on my mind, that it is the
“raffinose” of Loiseau (Compt. rend., 82, 1058). Whether the
Sugar is Berthelot’s mellitose as Tollens says raffinose is, I am not
prepared to say.

Scheibler can say whether my description of the crystals agrees
with his observations. My analytical numbers agree well with his,
for, although he burned the crystals, and I the dry substance, the

VOL. XLIX. G

74 ARMSTRONG AND MILLER: THE DECOMPOSITION AND

results obtained lead to the same formula. The optical activity
observed by me, sp. rt. pr. [a]; = 135°3°, agrees with the numbers
obtained by Scheibler and Tollens, [a]p = 104°; [a]; = 1147°;
they evidently, in this case also, have made the calculation for the
crystals, I for the dry substance: 114°7° for C,.H3.0,6,5H,0, is equal
to 135°1° for C\Hs.Oi.. The sp. gr. of a solution containing 1 gram
in 100 c.c. is higher than that given by Tollens; his figures (Ber., 18,
2616) leading to a true sp. gr. 1003712 for such a solution, my figure
is 1:003956. The yield of mucic acid is practically the same as
observed by Berthelot, Scheibler, and Tollens; my number is, however,
a little higher than that published by Scheibler, 30 per cent. against
26°7, but I cannot say with the quantity I had to work with that my
experiment can lay much claim to absolute accuracy. My observations
on the products of inversion do not.quite agree with those published
by Tollens. He says galactose, dextrose, and levulose are the pro-
ducts. I found, for the products of -the action of sulphuric acid,
[a]; = 43°5°, and K =:81; now, if we suppose that the remaining
19 non-K was unaltered sugar, and the 81 composed of equal
parts of galactose, dextrose, and levualose, we get a mixture the
optical activity of which works out [a]; = 36°1°, a figure sufficiently
removed from 43°5° to throw some doubt on the supposition; how-

ever, as I have not sufficient substance to settle the point at present, I
must leave it.

These facts, then, leave no doubt that I have been dealing with the
‘‘ raffinose ’’ of Loiseau, which Tollens says is Berthelot’s mellitose ;
the source whence I isolated*it seems to me of sufficient interest to

record.

X.—The decomposition and: genesis of .ydrocarbons at high tempera-
tures. I. The products.of the manufacture of gas from petroleum.

By Henry E. Armsrrone and A. K. Miter.

1. A PRELIMINARY account of.our work was given to the Society in
June of last year (Chem. News, 49, 285), a somewhat lengthy descrip-
tion of the manufacture of gas from oil, as practised in this country,
having been previously communicated by Dr. Armstrong to the
Society of Chemical Industry in a paper read before their London
Section in April, and printed in their Journal for September, 1884

(pp. 462—468).
2. The investigation was commenced several years ago, and was

GENESIS OF HYDROCARBONS AT HIGH TEMPERATURES. 75

carried on almost uninterruptedly during 1884 and the first half of
1885, but the products are so numerous and:their separation and iden-
tification is attended with such difficulty, that ‘the progress made is far
from satisfactory. Nevertheless, a stage has been reached at which
it appears desirable to record the results, as they are in part highly
suggestive and will lead te new inquiries: being undertaken ; besides
which the methods at our disposal are scarcely sufficient to enable us
to unravel the very tangled skein which the investigation presents.
An abstract of the present paper is given in the number of the
Abstracts of the Proceedings of the Chemical Society for June 18th, 1885.

3. The products of the manufacture of oil-gas are historically of great
interest, as their examination led Faraday, in 1825, to his momentous
discovery of “ Bicarburet of Hydrogen,” now. known as benzene.*
Another new compound described in the same paper was obtained
from the most volatile portien of the condensed liquor from the oil-
gas receivers. According to Faraday’s analysis, it contained carbon
and hydrogen in the same proportions as olefiant gas, but was of
double the density. This-is the composition -of. butylene, the dis-
covery of which has therefore always been attributed to Faraday; we
shall have occasion later on to consider whether the evidence he
adduces is sufficient to warrant this conclusion.

Faraday’s memoir affords a considerable amount of information
which we think justifies the inference that the oil-gas of the present
day differs but little from that then made,.although the oils now used
are derived from shale and petroleum and it was' then customary to
use fish or vegetable.oil.

4. We have not been able to ascertain thatiany-attempt has been
made to add to-our knowledge of the chemistry of the subject since
Faraday’s investigation was published, notwithstanding the attention
paid within recent years to-the manufacture of oil-gas, excepting that
Greville Williams appears to have commenced the..examination of the
bye-products of its manufacture not long after Dr. Armstrong had
begun to study them (Chem. News, 1884, p: 197; see also Jour. Soc.
Chem. Ind., 1884, p.-462).

5. In describing our work we propose to disregard the order in which
the various constituents were separated or identified; our main object
being to throw light on the nature of the changes resulting from the
decomposition of petroleum hydrocarbons at high temperatures, it
will be desirable to give .a connected account of the observations
relating to each of the series of hydrocarbons which occur in one
or other of the products examined, viz., the compressed gas itself ;

° ie On new compounds of carbon and hydrogen, and on certain other products
obtained during the decomposition of oil by heat :’’ a paper read June 16th, 1825,
by M. Faraday, F.R.S., &c., Phil. Trans., 1825, 440—466.

' G 2

76 ARMSTRONG AND MILLER: THE DECOMPOSITION AND

the liquid deposited during compression of the gas, either in a
chamber attached to the compressing pump or in the reservoir in
which the gas is stored; and the tar which is deposited from the
crude gas prior to compression. In this paper, however, we shall
deal only with that portion of the gas which is absorbed by bromine
and with the steam-distillable portion of the tar.

I. Benzenoid hydrocarbons.

6. The liquid deposited during compression of the gas is wholly
volatile in steam when recently obtained, and consists of hydrocarbons
capable of being polymerised by sulphuric acid, benzenes and a
relatively very small proportion of hydrocarbons unattackable by sul-
phuric acid; the last-mentioned constituents are present in larger
quantity in the steam distillate from the tar.

7. On mixing either liquid with moderately dilute sulphuric acid
(2 vols. acid 1 vol. water) much heat is developed, and by its action
the unsaturated hydrocarbons are for the most part converted into
compounds which are not distillable in steam; in order to effect as
complete a conversion as possible, it is desirable when the action
appears to be at an end to separate the acid from the oil and to treat
the latter with stronger acid—4.:1. In dealing with small quantities
a glass stoppered bottle is used; but for large quantities it is well to
use a copper can with a tightly fitting metal plug: there is then no
danger of the vessel bursting, and itis much easier to quickly cool the
mixture. . The acid should be added in small quantity at first, and
care should be taken in agitating; in working with large quantities
it is desirable to steam-distil as: soon as the treatment with 2 : 1 acid
is at an end, and to well mix the distillate with the stronger acid, as
owing to the viscid character of the product it becomes very difficult
to complete the conversion of the unsaturated hydrocarbons into
polymerides.

The steam distillate finally obtained has a peculiar characteristic
unpleasant odour, due apparently to the presence of volatile products
of the action of the acid.

8. A large quantity of benzene may be directly crystallised out from
the product so obtained from the liquid deposited from the gas during
compression in the manner first described by Faraday and made
popular many years later by Mansfield. We have employed a very
simple apparatus for the purpose, consisting of a cylindrical vessel,
closed only at the bottom, 12 inches high and 4°5 inches in diameter,
made of brass ,; of an inch thick; into this fits fairly closely a similar
cylinder, the bottom of which has numerous fine holes bored through
it. The larger cylinder is placed in a wooden pail or tub, surrounded

i, le ae

"oe ee ee ee ee eee

coer em HD

GENESIS OF HYDROCARBONS AT HIGH TEMPERATURES. 77

with a good freezing mixture, and the hydrocarbon is then poured
into it; the walls of the cylinder soon become coated with crystals,
which are detached by means of a stout metal rod, and this is fre-
quently used in stirring the contents of the cylinder. When crystal-
lisation is complete, the smaller cylinder is pushed down upon the
crystalline pulp, the pail is placed between the jaws of an ordinary
long carpenters’ cramp, a block of wood is placed across the mouth of
the inner cylinder, and pressure is gradually applied by turning the
screw of the cramp. The expressed liquid is syphoned off as it rises
into the interior of the cylinder. Pressure having been applied to a
sufficient extent and during a sufficient length of time, the screw is
released, the ram withdrawn, the cylinder lifted out of the freezing
mixture, and the small amount of liquid floating on the solid cake is
poured out; the benzene is then caused to melt by lowering into it a
tube fitted up after the manner of a wash-bottle, through which steam
is being passed.

If the mother-liquor from the crystals be fractionally distilled, and
the portions boiling near to 80° be then treated as above described, a
further considerable quantity of crystalline benzene may easily be
obtained.

In like manner the lowest fractions of the product obtained on
agitating the steam distillate from the tar with acid, &c., also yield
benzene.

The benzene thus separated is by no means pure, and should be
treated with alkaline permanganate, bromine, or sulphuric acid,
fractionally distilled, and again crystallised.

9. As much of the benzene as possible having been removed, the
residue is now extracted with hot concentrated sulphuric acid, and the
benzenes are recovered from the solution by hydrolysis (Chem. Soc.
Trans., 1884, 148).

The acid should be placed together with the hydrocarbon in a vessel
which can be closed—a stoppered bottle if a moderate quantity is to
be treated, or a copper can if the quantity be large. To ensure
intimate contact, the vessel is then very vigorously shaken. In the
first instance cold acid may be used, as heat is developed by the dis-
solution of the more easily attacked hydrocarbons; afterwards the
mixture of acid and hydrocarbon should be heated to 60—70°. An
insufficient quantity of acid is used in the first treatment, and the
unattacked hydrocarbon is afterwards shaken with fresh acid; finally
it is violently shaken with weakly fuming acid in order to remove all
the benzene. Care must be taken in doing this, as the action of the
fuming acid usually gives rise to the production of gas—sulphur
dioxide.

10. The residual unattacked hydrocarbon is then mixed with a little

ARMSTRONG AND MILLER: THE DECOMPOSITION AND

alkali and steam-distilled ;: the product is a brilliant colourless liquid,
having the pleasant sweet odour characteristic of a pure paraffin.

11. On hydrolysing: the. crude mixture of sulphonic acids, after the
whole of the benzenoid hydrocarbons have been recovered, a consider-
able amount of black carbonaceous matter remains with the acid; we
have always regarded this-as formed from products of the polymerisa-
tion of unsaturated hydrocarbons by the dilute acid during the first
treatment.

12. The method deseribed. is:that which renders it possible to sepa-
rate the whole of the benzenoid hydrocarbons from the original oil-gas
products. Much of the benzene, may, however, be obtained without
destroying the unsaturated hydrocarbons by refrigerating the appro-
priate fraction of the liquid deposited on compression of the gas or of
the steam distillate from the tar. And it is scarcely necessary to
point out, that either separate fractions of the original crude materials,
or the crude materials as a whole, may be submitted to.treatment in
the above manner.

13. To separate the various constituents of the complex mixture
of benzenoid hydrocarbons thus obtained, it-is fractionally distilled.
From the lowest fraction a considerable amount of benzene may easily
be frozen out. To separate the toluene, the fractions: boiling below
115°, from which no more benzene.can be obtained by refrigeration,
are shaken with hot concentrated sulphuric acid: as benzene is less
readily attacked’ than toiuene, it is possible by fractional treatment
with the acid to dissolve chiefly the latter, leaving a residue from
which a further quantity of benzene may be crystallised out. The
mixture of sulphonic acids with sulphuric acid is poured into water,
the solution neutralised with whiting paste, and the resulting calcium
salts are then converted into potassium salts, &c.; the liquid when
sufficiently concentrated deposits a large crop of the characteristic
crystals of potassium tolueneparasulphonate, which are recrystallised
and hydrolysed to recover the toluene.

14, The xylenes were separated by Jacobsen’s method (Ber., 10, 1009)
from the fractions boiling at 135—145°. Metaxrylene was identified by
conversion into its characteristic trinitro-derivative ; independent
evidence of its presence was afforded by the production of metatoluic
acid in large quantity on oxidation of the fractions collected between
135—145° of the original steam distillate from the oil-gas tar. Para-
aylene was obtained as sodium paraxylenesulphonate, C,Hy'SO;Na-H,0;
and paratoluic acid was obtained on oxidation of the hydrocarbon
separated by hydrolysis of this salt. Orthoxylene was separated in
the form of the highly characteristic sodium orthoxylenesulphonate,
C,HySO;Na‘5H,0. The three isomeric xylenes appear to us to
occur in the oil-gas product very much in the proportions in which

GENESIS OF HYDROCARBONS AT HIGH TEMPERATURES. 79

they ordinarily occur in coal-tar xylene, metaxylene being by far the
most abundant, and paraxylene the least abundant.

15. Mesitylene and pseudocumene were separated from the frac-
tions boiling between 155—175° by our method of fractional hydrolysis,
the hydrocarbon being for this purpose dissolved in sulphuric acid,
the solution diluted with water, and steam passed into it at a tempe-
rature not exceeding 100—105°. The hydrocarbon thus separated
was then reconverted into sulphonic acid; the barium salt of the acid
was almost pure mesitylenesulphonate, (C,H,,‘SO;),Ba*H,0, which is
a very characteristic salt. The equally characteristic potassium salt
was also prepared.

After separation of the main.bulk-of: the mesitylene, the tempera-
ture was raised, and the whole of the hydrocarbon present separated
by hydrolysis; it was then reeonverted into sulphonic acid, and the
latter several times recrystallised from: dilute sulphuric acid, as
recommended by Jacobsen (Annalen, 184, 198). In this way a consider-
able quantity of pure pseudocumenesulphonic acid was readily obtained.
The hydrocarbon separated from it had a constant boiling point, and
the trinitro- and tribromo-derivative prepared from.it had all the
properties of these derivatives of pseudocumene.

Mesitylene and pseudoeumene are, we think, also present in about
the same relative proportions in oil-gas tar as in coal-tar, but it is
much easier to separate them from.the former on aecount of the
absence of basic compounds.

16. The fractions 180—190° and 190—-200° were separately con-
verted into sulphonic acid, from which barium salts were prepared ; the
salts did not crystallise well, separating from a hot soldtion in an
apparently amorphous state. The dry salt was highly pulverulent,
and much like barium 1:2:3:5tetramethylbenzenesulphonate. The
salt obtained from the 180—190° fraction was found to contain 23°97,
24°02, 23°81 per cent. of barium. Theory indicates the presence of
24°33 per cent. barium in the sulphonate of a hydrocarbon of the
formula CH. The magnesium salt of the sulphonic acid prepared
from the 180—190° fraction crystallised well in flat needles, containing
24°6 per cent. water and 5°21 per cent. of magnesium; a salt of the
formula (CyHi3*SO;),.Mg*8H,0 should contain 24°24 per cent. water
and 5°33 per cent. magnesium. Magnesiam pseudocumenesulphonate
was prepared for comparison, and was found to be a very similar salt ;
it contained 26°32 per cent. of water and 5°56 per cent. magnesium.
A salt of the formula 2[{(C,H,,"SO,;),.Mg]‘17H.O would contain
26°61 per cent. water and 5°68 per cent. magnesium. A small
quantity of hydrocarbon separated from the barium salt prepared
from the 180—190° fraction boiled at 178—188°, chiefly at 180—
185°. It gave a solid bromo-derivative which, after recrystalli-

80 ARMSTRONG AND MILLER: THE DECOMPOSITION AND

sation from aleohdl mélted at 215°; after repeated recrystallisation
the melting point rose to 232°5°. The analysis of an impure portion
of this bromide gave 64°59 per cent. bromine, but only 0°09 gram of
substance could be used, and probably this contained tribromopseudo-
cumene; the amount corresponding with the formula ©,)H,,Br; is
64°69 per cent. bromine.

In the barium salt prepared from the 190—200° fraction 23°12
and 23°36 per cent..of barium was found ; a salt.of the formula
(Cy. H,s"SO;).Ba should contain 23°36 per cent. barium.

These facts leave little doubt that at least one benzene higher in
the series than trimethylbenzene is contained in the oil-gas product.
Probably isodurene is present, but unfortunately the amount at our
disposal was insufficient to enable us to prove this.*

17. A large quantity of naphthalene separates from the fractions
collected above 200°. After its removal a comparatively small amount
of hydrocarbon is left. Benzene being present.in such large amount
in the oil-gas product, it appeared not unlikely that diphenyl would
also occur in them. Notwithstanding persistent efforts to isolate this
hydrocarbon, however, we have entirely failed to discover it in the
portions boiling between 250—260°. Hitherto, indeed, we have been
unable to determine the nature of-any of the benzenoid hydrocarbons
of higher boiling point than naphthalene; the quantity obtained is

very small, and it is impossible to separate a pure substance by
distillation.

II. Hydrocarbons of the C,H2n—2 Series.

18. The separation and identification of the unsaturated hydro-
carbons other than those of the benzenoid series is compassed with diffi-
culty. By determining the amounts of bromine required to saturate the
various fractions, it was in the first instance ascertained that not only
olefines, but less saturated hydrocarbons were present. Moreover,
that there were no true acetylenes among these, that is, hydrocarbons
of the form CH-C-C,H»,:, at once appeared from the fact that
ammoniacal cuprous and argentic solutions were without appreciable
action upon the liquid condensed from the oil-gas. This conclusion
was confirmed by the observation made both at the Mansion House
Station Works of the Metropolitan District Railway Company and at
the Stratford Works of the Great Eastern Company that the gas

* Since this paper was placed in the printer’s hands, K. E. Schulze has announced
(Ber., 18, 3032) the discovery of 1:2:4:5tetramethylbenzene (durene) in coal-
tar. It may be added that, in a paper read at the last spring meeting of the Iron
and Steel Institute, Dr. Armstrong stated that the oil from the Jameson coke oven
contained benzenes different from those in ordinary coal-tar.

GENESIS UF HYDROCARBONS AT HIGH TEMPERATURES. 81

itself produced but a very small amount of precipitate in an ammo-
niacal cuprous solution.

19. On warming the crude liquid obtained from the reservoirs in
which the gas is stored, distillation at once sets in, much gas being at
first given off; this gas is absorbed by bromine. On steam-distilling
the resulting bromide, less than two-thirds passed over as a colourless
heavy oil; the residue became almost entirely solid on cooling, and
from it a pure substance was readily obtained by crystallisation from
alcohol.

In like manner‘by passing the compressed oil-gas into bromine, a
liquid bromide was obtained which was for the most part easily
volatile, only about 7 per cent. remaining when oil ceased to distil
over; the residue partially solidified on cooling, and from it more
than half its weight of a solid bromide was obtained identical with
that referred to in the previous paragraph.

The bromide in question erystallised from alcohol in small, glisten-
ing plates, exhibiting under:the microscope a very irregular outline ;
it fused at 116°, and volatilised with extreme slowness in a current of
steam. The percentage of bromine in it, as determined by Volhard’s
method of combustion with potassium.nitrate and sodium carbonate,
&c., was found to be 85°99. .Hence there could be little doubt that it
was a tetrabromide of the formula C,H,Br, the percentage of bromine
in which is 85°55. Judging from its properties, the bromide thus
obtained is identical with the crotonylene tetrabromide prepared by
Caventou from the liquid deposited on compressing coal-gas.

Theoretically, four distinct hydrocarbons of the formula C,H. are
possible, viz. :—

1. Ethylacetylene

2. Dimethylacetylene CMe:CMe.

3. Methylallene CHMe:C:CH,.

4, Vinylethylene (dimethyleneethane) CH,CH-’CH:CH:.

Ethylacetylene is excluded, as our hydrocarbon is not a true acetyl-
ene; the hydrocarbon obtained by Caventou from crude butylene
bromide gave a tetrabromide which volatilised somewhat readily in
the air, and from Almedingen’s experiments there is every reason to
suppose that this hydrocarbon was dimethylacetylene (Ber., 14,
2073). By distilling erythrol with formic acid, Heninger obtained a
hydrocarbon of the formula C,H, yielding a bromide similar to that
prepared by ourselves and apparently identical with that obtained by
Caventou by the decomposition of fusel oil at a red heat and from
coal-gas: as this hydrocarbon is formed from erythrol it may almost
certainly be regarded as dimethyleneethane or vinylethylene. On
inspection of the four formule it will be obvious that a study of the

82 ARMSTRONG AND MILLER: THE DECOMPOSITION AND

oxidation-products is calculated to afford the required proof: vinyl-
ethylene should not yield acetic acid, which would, however, be a
product of the oxidation of both dimethylacetylene and methylallene ;
and ethylacetylene forms propionic acid.

20. To obtain the hydrocarbon for oxidation, we have employed a
method which we believe will in the future be of great service in the
investigation of unsaturated hydrocarbons. The usual practice is to
withdraw the bromine from the bromides of these hydrocarbons by
means of sodium, but there are numerous objections to this method :
in many cases the change takes place only with difficulty and at a
high temperature ; in others secondary products are.formed owing to
the high temperature locally developed; and .the sedium usually
becomes coated with a protecting layer of bromide. It oceurred to
us that Gladstone and Tribe’s zinc-copper couple might be used with
advantage, they having already shown that ethylene and propylene
bromides are readily deprived of their bromine by its action in pre-
sence of alcohol (Chem. Soc. J., 1874, 406). The results have entirely
surpassed our expectations; as in all cases hitherto examined we
have obtained a practically theoretical yield of hydrocarbon by merely
warming the bromide with alcohol and the couple.. Moreover, the
hydrocarbon thus recovered has always been found to be identical
with that used in preparing the bromide—that is, it again yields the
same bromide. Unfortunately this method was not made use of until
nearly the close of our experiments; had we known of it earlier, we
feel sure that we should have been in a position to: throw far more
light on the nature of the products of the oil-gas manufacture. We
may add that it is our intention fully to inquire into the application
of this method to the separation of unsaturated hydrocarbons from
their compounds with halogens.

21. The oxidation of the hydrocarbon separated from the bromide
C,H,Br, by the action of the zinc-coppéer couple was effected by
displacing part of a 2 per cent. solution of potassium perman-
ganate from a bottle full of it by means of the gas from 50 grams
of the bromide—the gas having been collected over water in a
holder and left for some time in contact with the water to remove
alcohol-vapour. The permanganate was vigorously shaken to bring
it into contact with the gas, and the oxidation being completed,
sulphuric acid was added and the volatile acid removed by steam-
distillation. The distillate was neutralised by the addition of about
16 gram of sodium carbonate ; the neutral liquid having been con-
centrated, the required amount of silver nitrate was added, and the
solution boiled: an amount of silver was precipitated practically
equivalent to the silver nitrate used, and no trace of acetate could be
detected in the filtrate.

GENESIS OF HYDROCARBONS AT HIGH TEMPERATURES. 83

As formic acid is its oxidation-product, there can be no doubt that
the hydrocarbon of the formula C,H, from oil-gas is dimethylene-
ethane or vinylethylene, CH,,-CH’CH:CH2.

22. On steam-distilling the bromides prepared from the fractions
of the original liquid from the oil-gas reservoir collected within a few
degrees on either side of 40°, dark-coloured viscid residues were
obtained in which crystals gradually formed. It was easy to separate
these mechanically, and to purify them by recrystallisation from
alcohol. The pure substance crystallised in long, thin, narrow, well-
defined prisms, melting at 115°. Analysis gave results agreeing with
the formula C;H,Br,; thus :—

Subs. 0°1982 AgBr..0°3828 Bromine per cent. 82°19
0°2210 », 0°4284 82°49
0°2441 0°4734 : 82°53
0°2544 0°4935 : 82°55
0°2083 04021 - 82°15
0°2278 0 0°0471 Hydrogen 2°30
0°2612 0°0544 ” 2°31
0°2278 2 0°1322 Carbon 15°82

0°1526 ” 15°93

Percentages. Mean results.
15°47 15°87
2°07 2°30
82°46 82°38

100-00 100°55

23. The hydrocarbon was separated from this bromide by warming
it with the zinc-copper couple and alcohol; the amount obtained was
almost the theoretical. After several days’ digestion with calcium
chloride in a sealed tube it was distilled; it boiled almost constantly
at about 45°, leaving a few drops of a syrupy residue which exploded
when heated on platinum, in this respect behaving somewhat like
isoprene. It had the peculiar alliaceous odour so characteristic of the
crude liquid deposited from oil-gas. It was reconverted into the
original tetrabromide on careful treatment with bromine. Jt was
readily oxidised by a 4 per cent. solution of potassium permanganate :
more than half of the volatile acid produced was formic acid, the rest
being pure acetic acid.

Five hydrocarbons of the formula C;H, are at present known:
propylacetylene, isopropylacetylene, ethylmethylacetylene, piperylene
and isoprene ; ours appears to be a sixth. The first three of these
are excluded from consideration on account of their behaviour on

84 ARMSTRONG AND MILLER: THE DECOMPOSITION AND

oxidation ; moreover, our hydrocarbon is not a true acetylene, and
cannot, therefore, be either propyl- or isopropyl-acetylene. Piperylene,
according to Hofmann (Ber., 14, 665), forms a tetrabromide crystal-
lising from alcohol in glistening plates. Isoprene yields a liquid
bromide. Eight modifications in all are possible of a hydrocarbon of
the formula C;H,; three of these are derived from acetylene; the
formule of the remaining five are as follows :—

CH, CH; CH,
CH CH,
CH CH,
CH ¢
CH, CH,
Allylethylene. Isoallylethylene. Ethylallene. Symmetrical Unsymmetrical

dimethyl- dimethylallene.
allene.

Piperylene, according to Ladenburg (Ber., 16, 2059) is allylethylene ;
and that this is not the composition of our hydrocarbon may be
inferred from the fact that it yields acetic acid on oxidation. Ethy]-
allene is also excluded as it would furnish propionic acid on oxidation.
But three formule remain therefore, and we incline to select that of
isoallylethylene on account of the simple relation which this hydro-
carbon bears to normal amylene, from which probably our hydro-
carbon is immediately derived :—

CH;'CH,’CH,-CH-CH, = Propylethylene.
CH;'CH-‘CH:CH’CH, = Isoallylethylene.

If our argument be correct, isoprene must be either symmetrical or
unsymmetrical dimethylallene; taking into consideration all that is
known of the terpenes—to which it is so intimately related—the latter
is the more probable. We are now engaged in the study of the hydro-
carbons of the formula C,H, in the hope of solving this problem,
which is one of considerable importance in connection with the ques-
tion of the constitution of the hydrocarbons of the formula C,H.
24. A faulty determination of bromine, made by Volhard’s method,
led us, in the first instance, to regard the bromide just described
as identical with Schorlemmer’s so-called hexoylene tetrabromide,
C,H,Br,, whose description of this compound tallies very closely
with that which we have given of our tetrabromide. Schorlemmer
prepared his substance from a fraction boiling at about 80° of the
more volatile products of the distillation of boghead cannel} to this
he added bromine in excess, and after removing the greater part of
the admixed benzene by distillation, he heated the bromide with
sodium to regenerate the hydrocarbon; this was again brominated.

GENESIS OF HYDROCARBONS AT HIGH TEMPERATURES. 85

The product was an oil from which the solid gradually crystallised
out. It may be mentioned that Schorlemmer bases the formula
C,HiBr, on a single determination of bromine in only 0°1970 gram of
substance. In the hope, therefore, of obtaining hexoylene we very
carefully fractioned out a portion boiling at about 80—82° from our
crude material, and this was brominated; on steam-distilling the
product much benzene passed over, then an: oily bromide heavier
than water, a moderately limpid dark-brown oil remaining ; no crystals
whatever separated from this; and it was therefore submitted to the
action of the zinc-copper couple in presence of alcohol. The re-
generated hydrocarbon was again brominated and steam-distilled ;
very little came over. A minute quantity of solid was, however,
deposited in the condenser. The experiment was repeated with a
considerable quantity of the fractions collected at 70—80°: after
adding excess of bromine, the product was steam-distilled until
exhausted ; traces of a solid were again obtained. On:extracting the
residue with alcohol comparatively little dissolved and no crystals
could be obtained from the alcoholic extract: the final residue was a
carbonaceous mass. The solid referred to crystallised from alcohol
in small, hard, well-formed, short prisms melting at 185°; the
quantity obtained was too small even for an analysis.

25. In all cases, on steam-distilling the:crude bromides from the
various fractions, decomposition was observed to take place more or
less; and the non-volatile:residue from all but the lowest fractions
was more or less carbonaceous, and alcohol extracted but very little
oily matter, leaving a friable residue. The bromides which are thus
decomposed are probably derived from hydrocarbons of the C,H2»_2
series—or in part perhaps from less saturated hydrocarbons—such as
Schorlemmer has shown to be present in cannel oils. The polymer-
ides obtained by means of sulphuric acid were precisely of the
character of those described by Schorlemmer (Annalen, 1866, 139,
244),

III. Hydrocarbons of the Olefine group.

26. Indication of the presence of these hydrocarbons is afforded by
the behaviour with bromine especially of the lowest fractions of the
“hydrocarbon ”’ deposited during compression of oil-gas; these frac-
tions, if free from benzene, are entirely converted into polymerides and
soluble bodies on treatment with sulphuric acid: they therefore con-
sist of unsaturated hydrocarbons, but the amount of bromine which
they will absorb is far less than would be the case if they contained
only hydrocarbons less rich in hydrogen than the olefines.

To isolate the pure olefines from mixtures such as those with which

86 ARMSTRONG AND MILLER: THE DECOMPOSITION AND

we have had to deal, no ordinary method will suffice ; had we become
aware that the zinc-copper couple was applicable to the separation
of unsaturated hydrocarbons from their bromides at an early stage
instead of almost at the close of this portion of the investigation, it
would undoubtedly have been possible to isolate the olefines; but as
it; was we were obliged to content ourselves with the proof that ole-
fines were present, and with an indirect determination of their nature.

The method followed consisted in oxidising the various fractions by
agitation in a stoppered bottle with a cold 4 per cent. solution of
potassium permanganate. Jn selecting this method of treatment we
were guided by the knowledge that the normal olefines—those of the
type C,H2»+,CH*CH.—are converted by oxidation into acids of the
acetic series of the type C,H»,+,COOH, the CH, group being elimi-
nated as formic acid; only true acetylenes—i.e., hydrocarbons of the
formula C,H»+,C-CH yield similar products, and, as these were
known to be absent from our crude material, the production of the
corresponding acid from a fraction:of about the boiling point of any
particular normal olefine would be conclusive proof of the presence of
that olefine. Thus‘the normal amylene fraction should yield butyric
acid, the normal hexylene fraction valeric acid, &c.

The appropriate fractions having been oxidised, sulphuric acid was
added, and the volatile acid separated by steam-distillation ; the dis-
tillate was neutralised with sodium carbonate, concentrated, and then
fractionally precipitated with silver nitrate; the silver precipitates
were fractionally extracted with-water, the solntions well boiled to
decompose formate, and the dissolved salts crystallised out and
analysed. When a silver salt of constant composition-was obtained,
it was converted into the calcium salt, as the lower normal primary
acids of the acetic series all furnish characteristic calcium salts.

27. The amylene, hexylene, and heptylene fractions treated in this
way gave respectively normal ‘butyric, normal valerie, and normal
caproic acids.

The higher fractions were similarly treated, but repeated experi-
ments failed to yield any indication whatever of the presence of ole-
fines higher than heptylene: formic and acetic were the only volatile
paraffinoid acids produced—these being associated with the oxidation-
products of the benzenoid hydrocarbons present in the fractions
examined.

A very careful study of all the various fractions obtained has con-
vinced us that besides the three normal primary olefines above men-
tioned, no other hydrocarbons of the C,H2, series, or indeed of any
other paraffinoid series, can be present, except such as yield acetic
and formic acids on oxidation ; and we have no reason to suspect that
any olefines other than those mentioned are present in the liquid

GENESIS OF HYDROCARBONS AT HIGH TEMPERATURES. 87

products from the manufacture of oil-gas : our opinion being that the
acetic acid obtained was derived from hydrocarbons less saturated
than the olefines.

28. We have already referred to the presence of the hydrocarbon
C,H, in the compressed oil-gas (§ 19), and to the manner in which it
was separated from the crude mixture of bromides obtained on passing
the gas into bromine. On distilling the steam-distillate from this
mixture of bromides some hydrogen bromide was evolved, but after
a few distillations the constituent yielding this gas was practically all
decomposed. By far the largest amount of the bromide separated by
fractional distillation had about the boiling point of ethylene bromide,
and a considerable quantity of this-eompound was crystallised out
from this portion of the distillate by refrigeration.

29. The next largest fraction had about the boiling point of methyl-
ethylene (propylene) bromide, and'we have no doubt that it mainly
consisted of this compound, as acetic acid was obtained in large
quantity by direct oxidation of this fraction with permanganate.

30. Ethylene and propylene having thus been detected in the oil-gas,
and normal amylene, hexylene. and heptylene in the liquid deposited
from it, it was to be expected that normal butylene was also present.
A quantity of about 4000 grams of bromides from the gas gave,
however, but a relatively small quantity boiling at a higher tempera-
ture than propylene bromide, and as-it was impossible to separate a
pure product by distillation, the: various fraetions were directly
oxidised with permanganate. The quantities used were.:—

59 grams boiling at. 148—153°
42 ,, re 153—158
14 =, - 158—164
a 2 above 164°.

The acid distillate was treated with lead oxide in the manner recom-
mended by Linnemann (Annalen, 160, 222), in order to separate pro-
pionic acid, but this acid could not be detected; in fact, only acetic
and formic acids were formed.

We have before mentioned that a considerable quantity of bromides
was obtained by passing into bromine the gas given off when we
began to distil the liquid deposited from oil-gas on compression, and
that at least two-thirds of this was volatilised on steam-distillation,
the residue consisting mainly of crotonylene tetrabromide ; it is pos-
sible that the portion of the mixture of bromides volatile with water-
vapour contained butylene bromide, but most unfortunately the whole
of this material was lost in the fire which occurred in the laboratory
of the London Institution during the course of the investigation.

As butylene has a mach lower boiling, point than crotonylene, we

88 ARMSTRONG AND MILLER: THE DECOMPOSITION AND

should certainly expect to find its bromide among those obtained from
oil-gas itself, as a considerable quantity of crotonylene tetrabromide
is present; our failure to detect it has led us to consider the evidence
advanced by Faraday upon which the discovery of butylene among
the oil-gas products is attributed to him, and we are of opinion that
it is by no means conclusive. His words are as follows :—

“If a portion of the original liquid be warmed by the hand, or
otherwise, and the vapour which passes off be passed through a tube
at 0° (Fahr.), very little condensed vapour will go on to the mercurial
trough, but there will be found after a time a portion of fluid in the
tube distinguished by the following properties. Though a liquid at
0°, it upon slight elevation of temperature begins to boil, and before
it has attained 32° is all resolved into-vapour or gas. . . . . The
sp. gr. of the portion I obtained was between 27 and 28, hydrogen
being 1. . . . . When cooled to 0° it condensed again, and
inclosed in this state in a‘tube of known capacity and hermetically
sealed up, the bulk of a given weight of the substance at common
temperatures was ascertained. This compared with water gave the
sp. gr. of the liquid as 0°627 at 54°... . . . Alcohol dissolves it
in large quantity. ... . .. Sulphuric acid condenses the gas in very
large quantity: 1 volume of the acid condensing above 100 volumes
of the vapour. . . . . Great heat is produced during the action;
no sulphurous acid is formed ; the acid is much blackened, has a pecu-
liar odour, and upon dilution generally becomes turbid, but no gas
is evolved. A permanent compound of the acid with carbon and
hydrogen is produced, and enters as before mentioned into combina-
tion with bases. A mixture of 2 volumes of this vapour with 14 volumes
of pure oxygen was made, and a portion detonated in a eudiometer-
tube; 8°8 volumes of the mixture diminished by the spark to
5:7 volumes, and these by solution of potash to 14 volumes, which
were oxygen. Hence 7°4 volumes had been consumed, consisting

of :—

Vapour of substance ...........cccccecens

SIE 960000006006 betedsevsecseesaucce 63
Comtemic Geld formed, .2 cc cccccccccccccece 43
Oxygen in carbonic acid ..............006. 43
Oxygen combining with hydrogen.......... 2:0
Diminution by spark .............ceceeece 31

This is nearly as if 1 volume of the vapour or gas had required
6 volumes of oxygen, had consumed 4 of them in producing 4 of car-
bonic acid gas, and had occupied the other 2 by 40f hydrogen to form
water. Upon which view 4 volumes or proportionals of hydrogen
= 4, are combined with 4 proportionals of carbon = 24, to form

GENESIS OF HYDROCARBUNS AT HIGH TEMPERATURES. 89

1 volume of the vapour, the specific gravity of which would therefore
be 28. Now this is but little removed from the actual specific gravity
obtained by the preceding experiments; and knowing that this
vapour must contain small portions of other substances in solution,
there appears no reason to doubt that, if obtained pure, it would be
found thus constituted. . . . . Chlorine and the vapour were
therefore mixed in an exhausted retort: rapid combination took place,
much heat was evolved, and a liquor produced resembling hydro-
chloride of carbon, or the substance obtained by the same process
from olefiant gas. . . . . Further, it was composed of nearly
equal volumes of the vapour and chlorine; it could not, therefore, be
the same as the hydrochloride of carbon from olefiant gas, since it
contains twice as much carbon and hydrogen.”

Taking into account our own observations, especially the fact that
we have failed to detect butylene although crotonylene was obtained
in considerable quantity, we are inclined to think that Faraday was
the discoverer of crotonylene rather than of butylene. It is obvious
that he could not have had a pure substance for examination; the
ratio of the density of his gas to that of hydrogen was between 27 and
28, and these are numbers which represent the relative density of
crotonylene and butylene respectively ; the results of his combustion-
analysis are almost equally compatible with either formula, C,H, or
C,H,, always bearing in mind that probably both propylene and
amylene were present as impurities; we are not aware that the
behaviour of butylene with sulphuric acid has been studied, but it is
scarcely probable that it would be so readily absorbed as Faraday
describes; there remains but one fact which undonbtedly lends sup-
port to the conclusion that it was butylene, viz., that on mixing it
with chlorine a chloride was formed, ‘composed of nearly equal
volumes of the vapour and of chlorine.’ Our proof of the absence of
butylene, it should be added, holds good only on the assumption that
the butylene present is ethylethylene, as the two dimethylethylenes
would yield acetic acid on oxidation. As, however, all the olefines
which are proved to have been present are represented by the formula
C,Hn+:°CH-CHz, it does not appear probable that the butylene would
form an exception, especially as the crotonylene which is so abundantly
contained in oil-gas is indubitably a derivative of normal butylene.

IV. Hydrocarbons insoluble in sulphuric acid.

31. Our method of separating these hydrocarbons has already been
described (§§ 7—10). They are contained almost exclusively in the
portion boiling above 150° of the steam-distillate from oil-gas tar: for
example, about 1800 c.c. of fractions boiling at 105—130° gave after

VOL. XLIX. H

90 ARMSTRONG AND MILLER: THE DECOMPOSITION AND

exhaustive treatment with sulphuric acid only 6 grams of insoluble
hydrocarbons boiling at about 125—140°.

As far as possible, with the limited amount of material at our dis-
posal, the attempt was made to separate the mixture into its con-
stituents by fractional distillation ; but no very decided separation was
accomplished. Three fractions were analysed with the following
results :—

0°7637

0°7768

160—165°. 180—185°. 200—205°.
Carbon percentage ........ 85°33 85°37 85°27
. — 85°26 85°70
i _ 85°63 85°50
| -- a 85:39
ly Hydrogen percentage...... 14°80 14°74 14°32
= o #8 eeee — 14°73 —
Relative density, ~ ..... 0°7775 0°7886 0°8050

0°7980

” ” 20°

The general mean of the eight carbon determinations is 85°43 per
i cent., and of the four hydrogen determinations* 14°63 per cent. These
! numbers are very nearly those which correspond to the formula
} C,H», viz., 85°56 per cent. carbon and 14°32 per cent. hydrogen. A
| paraffin of the formula C,H, contains only 84°47 per cent. carbon and
15°53 per cent. hydrogen; while even that of the formula C,,H» con-
tains but 84°67 per cent. carbon and 15°33 per cent. hydrogen. The
conclusion that our products did not in the main consist of hydro-
carbons of the C,H:,+2 type is confirmed by the comparison of our
determinations of relative density with the data given by Krafft
( Ber., 15, 1687) for the normal paraffins :—

B.p. Density at 0°. Density at 20°.
eS Serre 173°0° 0°7452 gram. 0°7304 gram.
CyHoy ....---- 194°5 0°7557 ss, O-7411 ,,
214°5 0°7655—,, 07511 _ ,,

Our figures are in every case considerably higher, and the difference
would be greater if a strict comparison were made by calculating the
densities corresponding to our relative densities. We are satisfied
that this is not due to the presence of benzevoid hydrocarbons, as
special care was taken to remove these by treatment with fuming
sulphuric and nitric acids followed by distillation from sodium.

* As a warning to those who, like myself, are in the habit of using compressed
oxygen for combustions, I may mention that several hydrogen determinations were
lost, owing, as was afterwards discovered, to the presence of traces of hydrogen in

the oxygen.—H. E. A.

GENESIS OF HYDROCARBONS AT HIGH TEMPERATURES. 91

Our figures more nearly agree with those given by Markownikoff
and Oglobine (Ann. Chim. Phys., 1884 [6], 2,372) for the C,H» hydro-
carbons which they separated—probably in a state of only approximate
purity—from Russian petroleum :—

B. p. Density at 0°.
161° 0°795 gram.
180 08119 _ ,,
196 08025,

Hence we are of opinion that the portion insoluble in sulphuric acid
of the steam-distillate from oil-gas tar which we have examined con-
tained both true paraffins and pseudolefines such as mainly compose
Russian petroleum, the latter being probably the principal con-
stituents.

V. Summary and discussion of results.

32. Thus far we have been led to recognise among the products of
the manufacture of oil-gas the following hydrocarbons :—

a. Paraffins, of which traces only may be said to be present.

b. ““ Pseudolefines,” that is, saturated hydrocarbons of the C,H.»
series such as occur in Russian petroleum; these also are present in
relatively small amount.

c. Olefines, viz., ethylene, propylene, normal amylene, normal
hexylene and normal heptylene, all higher homologues being absent.
Ethylene is an important constituent of oil-gas as used, and so also
apparently is propylene. The liquid deposited from the crude gas on
compression is moderately rich in amylene, hexylene and heptylene.

d. “ Pseudacetylenes,” viz.,. crotonylene (dimethyleneethane),
CH,CH-CH-CH;, and isoallylethylene, CH;CH-CH-CH-CH;. The
former is probably an important constituent of the gas, being of high
value as an illuminant. Besides these two, both the liquid deposited
on compression of the gas and the tar are rich in hydrocarbons
identical with, or very closely related to, those discovered by Schor-
lemmer in cannel oils.

e. Benzenoid hydrocarbons, viz., benzene, toluene, the three isomeric
dimethylbenzenes, the two trimethylbenzenes—pseudocumene and
mesitylene—and naphthalene; the first mentioned is a specially im-
portant constituent in point of quantity. There is reason to believe
that other benzenoid hydrocarbons besides these are present, even in
the portions of the tar volatile with steam.

33. We have had the opportunity of examining various samples of
the oil—shale oil or crude petroleam—used at the works from which
we have obtained our materials, and have satisfied ourselves that the

H 2

FE
=
PS
Lt
¢.

¢ "<
a
“4
&
>
2 ;

-
* &

92 ARMSTRONG AND MILLER: THE DECOMPOSITION AND

proportion of constituents in them volatile with water-vapour was small,
and that they were of a different character from those met with in
the bye-products from the manufacture; we have therefore little
doubt that practically all the above-mentioned substances are pro-
duced in the course of the manufacture of the gas.

34. The paraffins are probably formed in the manner indicated by
Thorpe and Young (Proc. Roy. Soc., 1873, 21, 184), although it is pos-
sible that they are in part original constituents of the oil used. It is,
however, noteworthy that whereas Thorpe and Young in their inves-
tigation of the decomposition of solid paraffins by heat obtained
mixtures of lower paraffins and olefines almost in equimolecular pro-
portions from pentane upwards, our material, although rich in
amylene and its next two homologues, has not been found to contain
the corresponding paraffins.

35. The “ pseudolefines ” are also more probably products of change
than original constituents. But it is to be remembered that accord-
ing to Beilstein and Kurbatow (Ber., 13, 2028), American petroleum
eontains hydrocarbons of the C,,H:, series similar to those in Russian
petroleum: hence, assuming our view to be correct, the question
arises whether the pseudolefines are formed by simplification of higher
pseudolefines or by the removal of hydrogen from corresponding
paraffins. Theoretically this question is one of considerable import-
ance, and we are therefore making it the subject of special experi-
mental study.

36. The presence of olefines in products of the distillation of cannel,
of coal and of paraffin, has been established by various observers, but
no proof of their nature had hitherto been given; it is therefore a
matter of interest that those detected by us are all of the type
C, Hen, 1°CH-CH,.

37. Regarding the hydrocarbons of the C,H2n_: series, it is to be noted
that vinylethylene is present in much larger proportion than isoallyl-
ethylene, and hence it may be inferred that the former is a much
more stable compound. Judging from the behaviour on oxidation of
the remaining unsaturated hydrocarbons, which it is to be assumed
are to a large extent also members of the C,H2»_, series, it appears
probable that they are formed from corresponding normal paraffins by
processes similar to those by which these paraffins are converted into
normal olefines and by which vinylethylene and isoallylethylene are
formed from the corresponding normal olefines or paraffins.

38. Since Berthelot’s discovery of the formation of benzene from
acetylene, the benzenoid hydrocarbons have always been regarded as
built up from true acetylenes (comp. Jacobsen, Ber., 10, 853). The
fact that true acetylenes are all but absent from oil-gas and the bye-
products of its manufacture would lead us to doubt whether this is

a a a ae a

GENESIS OF HYDROCARBONS AT HIGH TEMPERATURES. 93

so entirely the case; and the question arises whether these hydro-
carbons may not also be directly descended from corresponding
paraffins: whether, for example, benzene may not be obtainable
directly from hexane by withdrawal of hydrogen. It is sufficient to
point this out, and it would be useless to further discuss the question ;
but we shall endeavour to solve it by experiment.

It will suffice to have pointed out thus briefly the various problems
which await experimental investigation arising out of the examina-
tion of the products of the decomposition of petroleum hydrocarbons
in the manufacture of oil-gas. By continuing the examination of
these products, and also by the study of the changes undergone by a
material of more definite composition than the oils used for the pur-
pose, viz., solid paraffin, we hope to obtain facts which will serve to
elucidate the nature of many of the changes in hydrocarbons which
occur at higher temperatures. The subject is one of very great
importance, both as bearing on the discovery of rational methods of
coking coal and of manufacturing illuminating gas and also hydro-
carbons such as benzene, naphthalene and anthracene. It has
undoubtedly also a bearing upon the question as to the origin of the
complex mixtures of hydrocarbons such as constitute the different
varieties of natural petroleum.

Dr. Armstrong has in his previous paper expressed his thanks to
various gentlemen connected with the Great Eastern and Metro-
politan District Railways for aid afforded him in carrying out this
investigation. We desire on this occasion, however, to thank
Mr. W. F. Pettigrew, Engineer in charge of the Great Eastern Com-
paay’s oil-gas works, for his ever-ready assistance during the past
year.

City and Guilds of London Institute,
Central Institution,
December, 1885.

XI.—The Combustion of Carbonic Oxide and Hydrogen.

By Harotp Drxon, M.A., the Duke of Bedford’s Lecturer in
Chemistry, Balliol College, Oxford.

Part I.—Tue Action or Stream on Carnonic OXIDE.

In a preliminary account of a research on the incomplete combustion
of carbonic oxide and hydrogen, read in 1880 before the British
Association, I published the fact that a mixture of carefully dried
carbonic oxide and oxygen would not explode when electric sparks
were passed through it, but that on the addition of a minute trace of
water, or volatile body containing hydrogen, the mixture became
inflammable.

To account for this fact, I ventured, after making many experiments,
to put forward the hypothesis that, in the ordinary combustion of
carbonic oxide, the steam present acts the part of a “carrier of
oxygen” by undergoing reductions and successive re-formations :—

(1.) CO + H,O = CO, +H,
(2.) 2H, + O. = 2H.0.

The influence of steam in determining the explosion of carbonic
oxide and oxygen has since been confirmed by several experimenters ;
but, as some* discussion has arisen as to the mode in which the steam
exerts its influence, I venture to bring before the Society the reasons
which appear to me to show that the hypothesis referred to gives the
simplest explanation of the observed facts.

Starting on the assumption that dry carbonic oxide and oxygen do
not explode, and that the presence of a minute trace of steam confers
inflammability on the mixture, I will examine some of the methods by
which the steam may be suppesed to exert its influence.

The first question to be asked is—Does the steam act in virtue
of some property peculiar to itself, or does it share this catalytic
power with other gases? To answer this question, the following
experiments were made:—Small quantities of different gases were
added to separate portions of a non-inflammable mixture of dry
carbonic oxide and oxygen, and the spark was then passed. Inall
cases where a gas containing hydrogen was introduced, the mixture
exploded ; in all cases where a gas containing no hydrogen was it
troduced, the mixture did not explode. This argument by the Jowt

* See Prof. Armstrong’s address to Chemical Section of British Association s#
Aberdeen.

DIXON: COMBUSTION OF CARBONIC OXIDE AND HYDROGEN. 95

Method becomes stronger the greater the number of instances, positive
and negative, adduced, and the nearer the positive and negative
instances are in kind to each other. In the following table, under
“positive instances” is given the composition of those gases, the
presence of which determined the explosion of dry carbonic oxide and
oxygen; under “negative instances” is given the composition of

those gases which did not determine the explosion :—

Positive instances. Negative instances.

Thus whilst an oxide of hydrogen and sulphide of hydrogen both
determined the explosion, a third gas composed of sulphur and
oxygen did not; whilst a sulphide of hydrogen and a carbide of
hydrogen both determined the explosion, a third gas composed of
sulphur and carbon did not. Oxygen, sulphur, carbon, nitrogen, and
chlorine when combined one with the other, had no action; the same
elements combined with hydrogen determined the explosion. Thus
only steam—or bodies which form steam under the conditions of the
experiment—were found capable of determining the explosion; other
gaseous bodies, which differ from the former in not forming steam,
were found to have no action. The conclusion is irresistible that the
steam does not act as a mere “ third body,” but, in virtue of its own
peculiar chemical properties.

It being shown, then, that steam determines the explosion of car-
bonic oxide and oxygen in virtue of its chemical properties, the second
question is—Does the steam act in some unknown way, by its presence,
without undergoing chemical change, or does it undergo a cycle of
chemical reactions whereby it gives up oxygen to the carbonic oxide,
and returns to its original state? To this question, I believe, no

96 DIXON: THE COMBUSTION OF

absolute answer can be given with our present knowledge, for we have
no direct evidence that the steam is chemically changed during the
explosion. But if it can be shown that such a cycle of chemical
reactions can take place under the conditions of the experiment, then
it is reasonable to prefer an explanation which accounts for the action
by known forces, to an explanation which involves the existence of
unknown forces. Such a cycle can occur, and, as I shall show later
on, there is strong experimental evidence against the mere “ contact ”
action of steam.

What then are the reactions by which steam can convey oxygen to
carbonic oxide? My hypothesis is that the carbonic oxide is oxidised
by the steam in the path of the spark, and that the hydrogen set free
unites with oxygen to form steam at a high temperature; the steam
so produced acts on more carbonic oxide, and so on, until all the
oxygen or carbonic oxide is used up. It is necessary for this complex
change that carbonic oxide should decompose steam at the temperature
reached in the explosion, and that hydrogen should unite with oxygen
under the same conditions. These two reactions are fundamental.
If either of them is proved not to occur under the conditions of the
experiment, the hypothesis falls to the ground. Ina memoir “ On
the Conditions of Chemical Change in Gases” (Phil. Trans., 1884),
I brought forward experiments which appeared to me to establish
these two reactions: for (1) when carbonic oxide was exploded with
defect of oxygen in a moistened eudiometer, the carbonic acid formed
was found to be more than twice the oxygen used, and a correspond-
ing quantity of free hydrogen was found to be present in the residue ;
and (2) in a dry mixture of hydrogen and oxygen an electric spark
caused an explosion under a pressure exactly as low as in a damp
mixture. But in a recent paper read before the German Chemical
Society, Moritz Traube has rejected my explanation on the ground
that carbonic oxide does not decompose steam at a high temperature ;
and he has put forward the hypothesis that the steam determines the
explosion by a joint action with the carbonic oxide on the oxygen,
whereby the steam is oxidised to hydrogen peroxide, and the carbonic
oxide to carbonic acid :—

(1.) CO + H,O + O, = CO, + H,0.,,
and that the peroxide of hydrogen is then reduced to steam by more
carbonic oxide :—

(2.) co + H,0, — co, + H,0.

Moritz Traube says :—
“ When moist carbonic oxide is submitted to the action of induction
sparks for several hours, its volume is not altered, and no trace

CARBONIC OXIDE AND HYDROGEN. 97

of carbonic acid and hydrogen are formed. In opposition to
Dixon’s supposition, hydrogen has the reverse action on carbonic
acid, reducing it at a white heat; for if electric sparks are con-
tinuously passed through a mixture of carbonic acid and hydro-
gen, carbonic oxide and steam are formed with a corresponding
diminution of volume.”

In these positive statements, Traube asserts not only that carbonic
oxide does not decompose steam at the temperature caused by the
passage of the induction spark, but he argues that the reaction is
impossible because the reverse action occurs under the same con-
ditions. I will first deal with these two statements, and afterwards
discuss his own explanation.

1. The Action of Steam on Carbonic Oxide in contact with Red-hot
Platinum.

The oxidation of carbonic oxide by steam at a high temperature was
discovered by Grove. In the paper read before the Royal Society in
1846, “ On certain Phenomena of Voltaic Ignition and the Decompo-
sition of Water by Heat,” which formed the Bakerian Lecture for
1847, Grove describes two experiments which completely refute
Traube’s statements. These two experiments seem to have been
obscured by the brilliant discovery to which they led—the decomposi-
tion of water at a high temperature. I can find no mention of them
in any of the chemical manuals I have consulted. Bunsen when he
made his experiments on the incomplete combustion of gases, was
unaware of the oxidation of carbonic oxide by steam in an explosion.
In 1876 the reaction was rediscovered independently by Horstmann
and myself in repeating Bunsen’s experiments. In the work I have
lately been engaged in—viz., heating a coil of platinum wire first in
carbonic oxide and steam, and then in carbonic acid and hydrogen,
Ihave been unconsciously repeating the experiments made by Grove
40 years ago. Employing the current from the nitric acid battery
he had invented a few years previously, Grove heated a loop of
platinum wire sealed into a glass eudiometer containing various
gases :—

“Hydrogen and carbonic acid mixed in equal volumes were readily
acted on by the ignited wire; they contracted to 0°48 of the
original volume; the residue was carbonic oxide; one equivalent
of oxygen had therefore united with the hydrogen.

“Carbonic oxide exhibited a remarkable effect, and one which,
coupled with the last experiment, gave rise to considerations
which mainly led to the results to be detailed in the body of

DIXON: THE COMBUSTION OF

this paper. Carbonic oxide, very pure, and carefully freed from
carbonic acid, was exposed to the ignited wire over distilled water;
the gas increased in volume in one experiment to one-third of its
original volume—in the greater number of instances to one-fifth :
this increase depended upon the intensity of ignition, which it
was very difficult to maintain at its maximum on account of
the frequent fusions of the platinum wires.

“Here again I had a long research—and many erroneous guesses—
which I need not detail. The effect did not take place with
perfectly dry gas over mercury, and I thence was led to attribute
it to some combination with aqueous vapour ; the increase turned
out to be occasioned by the formation of carbonic acid. By agita-
tion with caustic potash or lime-water the gas was reduced to
exactly its original bulk, but it was now found to be mixed
with a volume of hydrogen equal to the volume of carbonic acid
by which it had been increased ; it was thus perfectly clear that
half a volume of oxygen derived from the vapour of the water
had combined with one volume of carbonic oxide, and formed one
volume of carbonic acid, leaving in place of the carbonic oxide
with which it had combined the one volume of hydrogen with
which it had been originally associated.

“Comparing the last experiment, viz., that of mixed carbonic acid
and hydrogen with this, I was naturally struck with the curious
reversal of affinities under circumstances so nearly similar; in
the one case hydrogen taking oxygen from carbonic acid to
form water, and leaving carbonic oxide; in the other, carbonic
oxide taking oxygen from water to form carbonic acid and
leaving hydrogen.”

This “ curious reversal of affinities” led Grove to test the stability
of steam at a high temperature—an investigation which is classical in
chemistry. The facts themselves are explained by the “ Law of
Mass.” When carbonic oxide and steam are heated they react to
form carbonic acid and hydrogen; the carbonic acid and hydrogen so
produced react to form carbonic oxide and steam. At a certain point
in the decomposition (depending on the conditions of the experiment)
the rate of decomposition is equalled by the rate of recomposition ;
the mixture then remains in equilibrium. So that if a platinum wire
is heated in a mixture of carbonic oxide and steam, carbonic acid and
hydrogen are produced. As the carbonic oxide and steam are ex-
hausted, the rate of this change becomes slower ; but as the carbonic
acid and hydrogen increase, the rate of the inverse change becomes
quicker. An equilibrium is therefore reached when the two inverse
changes take place at equal rates, and it is immaterial which pair of

CARBONIC OXIDE AND HYTORGEN. 99

gases are present to start with ; the same equilifv&jm is reached when
carbonic oxide is heated in steam, or carboni jd, is heated in
hydrogen.

The carbonic oxide employed in the following ex POengnts was
prepared from anhydrous magnesium formate; the coal cid was
prepared by the action of dilute hydric sulphate on poder 3and
spar; the oxygen by heating well-dried potassium chlorat the
hydrogen by electrolysis of dilute hydric sulphate. The gases
collected over mercury and dried by anhydrous phosphoric acid.

Into a straight glass eudiometer two coils of platinum were attached
to thick platinum pieces, AA’, BB’, sealed into the glass. A current
from three Grove cells served to heat both coils to bright redness.
A small ring of gas-jets, C, surrounding the eudiometer served to
keep hot the top of the mercury column and the glass between the
mercury and the coils.

In the first experiment the current was diminished until the coils
were just invisible in the dark. The volume of carbonic oxide and
steam remained unaltered. When the coils were visible in the dark
a very slight expansion was noticed. On heating the coils to dull
redness the volume gradually increased and became constant after
four hours. On passing up into the tube a few drops of a dilute

solution of potash the gases contracted to their original volume.

Expt. 1. Pressure 600 mm.
Platinum Coils heated in Carbonic Oxide and Steam to Dull Redness.
Volume of dry gas reduced to 0° and 760 mm. Percentage.
Vol. of carbonic oxide at starting .. 22°20 c.c. 100-0
After heating for 1 hour 22°55 101°6
22°91 103°2
- 23°35 105°2
- a 3 23°35 105°2
After adding potash 22°24 100°2
This experiment shows that out of 100 volumes of carbonic oxide,
52 volumes were converted into carbonic acid with the liberation of
5°2 volumes of hydrogen.

Expt. 2. Pressure 600 mm.

Platinum Coils heated in Carbonic Oxide and Steam to Redness.

”?

Volume of dry gas reduced to 0° and 760 mm. Percentage.
Vol. of carbonic oxide at starting .. 23°87 c.c. 100°0
After heating for 1 hour “Ot 107°0

- » 4 hours ,° 111°7
” ~» © 6 111°6
After adding potash 100°0

DIXON: THE COMBUSTION OF

Eudiometer with Platinum Coils.

Expt. 3. Pressure 600 mm.
Platinum Coils heated in Carbonic Oxide and Steam to Bright Redness.

Volume of dry gas reduced to 0° and 760 mm. Percentage.
Vol. of carbonic oxide at starting .. 24°94 ¢.c. 100°0
After heating 4 hours 02 1143
After adding potash 4: 9 100°1

These experiments show that steam begins to oxidise carbonic oxide
at a dull red heat, and that at a bright red heat about one-sixth the
carbonic oxide is converted into carbonic acid.

CARBONIC OXIDE AND HYDROGEN. 101

The Action of Hydrogen on Carbonic Acid in contact with Red-hot
Platinum.

In a second series of experiments, equal volumes of dry carbonic
acid and hydrogen were brought into the eudiometer, and the two
coils were heated in the same way as before.

Platinum Coils heated in Carbonic Acid and Hydrogen to Bright
Redness.

Volume of dry gas reduced to 0° C. and 760 mm. Percentage.
Vol. of carbonic acid . 1000
” 100°0
200°0
After heating for 1 hour ; 130°7
SS eee ' 1140
” 12 ’ : 113°6

”

On heating the coils, the hydrogen reduces the carbonic acid with
the formation of steam and carbonic oxide, the volume of the dry
gas therefore diminishes.

In this experiment, in which carbonic acid and hydrogen were
taken to start with, the final equilibrium was practically the same as
in the previous experiment in which the initial gases were carbonic
oxide and steam.

The limited reduction of carbonic acid by hydrogen in presence of the
products of their reaction serves to confirm, instead of to disprove,
the oxidation of carbonic oxide by steam.

In both these experiments, it is possible to push the reaction
farther by removing one of the products of the change. For instance
when steam acts on carbonic oxide and the carbonic acid is removed
as it is formed, the whole of the carbonic oxide can be oxidised ; and,
conversely, when hydrogen acts on carbonic acid and the steam is
removed as it is formed, the whole of the carbonic acid can be reduced.

3. Action of Steam on Carbonic Oxide in presence of Potash.

Three similar eudiometers, each fitted with a single coil of
platinum wire, were charged with an equal volume of carbonic oxide.
Some dilute solution of potash was introduced into the three tubes,
and the wires were heated. After four hours’ heating, the gases in
the first tube were analysed ; after 16 hours’ heating, the gases in the
second tube were analysed; and after 64 hours’ heating, the gases in
the third tube were analysed.

102 ' DIXON: THE COMBUSTION OF

1. After 4 hours’ 2. After16 hours’ 3. After 64 hours’
heating. heating. heating.
Carbonic oxide .. 75°2 39°5 29°9
60°5 70°1

100°0 100°0

In a second experiment, a eudiometer with two coils was used, and
the wires were kept hotter than in the previous series. After four
hours’ heating, one-third of the gases was removed, and the remainder
was heated for 12 hours longer. Then half this remainder was
removed, and the residue was heated for 16 hours longer.

1. After 4 hours’ 2. After 16 hours’ 3. After 32 hours’
heating. heating. heating.

Carbonic oxide .. 15°0 2°8 08
99°2

100°0

These experiments show that, when the carbonic acid formed is
removed from the sphere of action, carbonic oxide may be entirely
oxidised by excess of steam.

4. Action of Hydrogen on Carbonic Acid in presence of Anhydrous
Phosphoric Acid.

Dry carbonic acid was mixed with an excess of dry hydrogen in the
eudiometer, and several short sticks of anhydrous phosphoric acid
were introduced into the gases. After the volume of the gases had
been measured, the coils were heated to redness. From time to time
the eudiometer was cooled and the contraction observed. After
heating for 40 hours, all but 0°4 per cent. of the carbonic acid had

been reduced.

Platinum Coils heated to Redness in Carbonic Acid and Hydrogen in
presence of P,O;.

Volume of dry gas reduced to 0° and 760 mm. Percentage.
Vol. of carbonic acid - 1000
Vol. of hydrogen ; 154°6

254°6
155°0

99°6

CARBONIC OXIDE AND HYDROGEN. 103

5. The Action of Induction Sparks on Carbonic Oxide and Steam.

The fact that steam oxidises carbonic oxide when induction sparks
are passed through the mixed gases was discovered by Buff and
Hofmann,* who also confirmed Grove’s observation that an incan-
descent platinum wire transforms carbonic oxide, when confined
over water, into carbonic acid. Buff and Hofmann give no numbers.
In repeating these experiments, sparks from an induction coil were
passed through carbonic oxide and steam between two spheres of
platinum-iridium 2 mm. apart. The volume increased rapidly at
first, but after an hour remained nearly constant. On adding potash,
after 12 hours’ sparking, the gases contracted to less than their
original volume, showing that some other gas besides carbonic acid
had been formed. The solution was tested, and was found to give all
the reactions of a formate.

Volume of dry gas reduced to 0° and 760 mm. Percentage.
Vol. of carbonic oxide at starting .. 18°80 c.c. 100°0
After L hour sparking 9: ie 103°6
After 12 hours’ sparking : " 1043s
After adding potash , 83°9

In another experiment, a large Leyden jar was connected with the
secondary coil of the Ruhmkorff, and very powerful sparks were passed
from this apparatus through the mixture of carbonic oxide and steam.
The mixture expanded to about the same extent as before, but the glass
near the wires was covered with a thick black deposit, which proved
to be nearly all carbon. On heating the deposit in the air it glowed
and disappeared, leaving a small grey metallic residue, no doubt
volatilised from the wires. When potash solution was added to the
mixture, the gases, as before, contracted to less than their original
volume; but the solution when tested showed that not more than a
trace of formate was present. Some carbonic oxide had, therefore,
been decomposed with the formation of carbon and carbonic acid. It
appears from this experiment that the decomposition of carbonic
oxide, which quickly reaches its limit in the dry gas,t takes place to
a much larger extent when the liberated oxygen is removed by union
with undecomposed carbonic oxide in presence of steam.

* “On the Decomposition of Gaseous Compounds.” Buff and Hofmann (J.
Chem. Soc., 12, 282).

+ When sparks from a Leyden jar were sent through dry carbonic oxide, a slight
decomposition took place amounting to about 0°5 per cent. of the original gas.

DIXON: THE COMBUSTION OF

Powerful Sparks through Carbonic Oxide and Steam.

Volume reduced to 0° and 760 mm. Percentage.
Vol. of carbonic oxide at starting .. 22°66 c.c. 100°0
After 16 hours’ sparking 3° 104°7
After addition of potash 4. 85°8

In another experiment, sparks from the Leyden jar were passed for
one hour through carbonic oxide and steam; the volume expanded
about 4 per cent., but no trace of formate could be detected in the

tube.

6. The Actionof Induction Sparks on Carbonic Acid and Hydrogen.

When induction sparks were passed through a mixture of carbonic
acid and hydrogen, in which the hydrogen was in slight excess, the
volume quickly contracted, and after three hours only 6°4 per cent. of
the carbonic acid remained undecomposed, as shown by the contrac-
tion on adding potash. A trace of formate was found in the solution,
so that the carbonic acid remaining was probably less than 6 per cent.
The sparking caused a very slight yellow deposit on the glass near the
wires.

Volumes reduced to 0° C. and 760 mm. Percentage.
Vol. of carbonic acid at starting .. 22°86 c.c. 100°0
Vol. of hydrogen at starting 23°37

46°23
Vol. after 1 hour sparking 25°05
3 hours’ sparking 24°70
adding potash .......... 23°24

Contraction on adding potash......

”

”?

7. The Incomplete Combustion of Carbonic Oxide in presence of varying
Quantities of Steam.

At the beginning of this paper, [ mentioned an experiment which
showed that when carbonic acid was exploded with insufficient oxygen
to completely burn it in a moist eudiometer, the excess of carbonic
oxide reacted with the steam at the high temperature produced to
form carbonic acid and hydrogen. With the small quantity of steam
present at ordinary temperatures, only a small quantity of free
hydrogen is so produced ; but by raising the temperature of the eudio-
meter the proportion of steam can be increased at pleasure, and @
proportionately large quantity of hydrogen is then found in the
residue. A mixture of carbonic oxide and oxygen was made in the

CARBONIC OXIDE AND HYDROGEN. 105

proportion of about 3 to 1. Half of this mixture was fired in a
eudiometer saturated with steam at 52° C.; the second half of the
mixture was fired under the same pressure, but saturated with steam
at a temperature of 68°. The number of steam molecules in the
first case was half that of the oxygen; in the second case the number
of steam molecules present was the same as that of the oxygen. In
the following table the results of these experiments are given, together
with those of the original experiment at the ordinary temperature.

Expt. 1.
At 15°. Tension of Steam 13 mm.

Before the explosion. After the explosion.

0. H,0. . | ©0.

10?

Analysis of Residue...

Found.

Expt. 2.
At 52°. Tension of Steam 102 mmv-

Before the explosion. After the explosion.

O>. H,0. CO,. co.

50 211°4 | 85 *4

Analysis of Residue.

| Found. Calculated.

Micamakistalnoens

VOL. XLIX.

DIXON: THE OOMBUSTION OF

Expt. 3.
At 68°. Tension of Steam 214 mm.

Before the explosion. After the explosion.

03. H.0.

100 100

Analysis of Residue.

Found. | Calculated.

GAP 00cencessececane ‘ 5°7
i stupebsiendane 1‘1

Although some of the water on the sides of the eudiometer was no
doubt evaporated during these explosions, and took part in the
reaction, the experiments bring out clearly the fact that as the
quantity of steam is increased the quantity of carbonic oxide oxidised
by it increases also. In the last experiment more than 10 per cent. of
the carbonic acid formed was owing to the action of the steam on
the carbonic oxide.

It has thus, I think, been amply proved by different experimenters
that when steam and carbonic oxide are heated together, either by an
induction spark or by a red-hot coil of platinum, or by heated gases
in their neighbourhood—the steam oxidises some of the carbonic
oxide ; and, if fresh proof were required, it has been afforded in a paper
(Ber., 1885, 2894) just published by Naumann and Pistor. Since my
experiments were made, these chemists have shown that when
carbonic oxide and steam are heated in a tube below 580° no reaction
takes place ; a little above 600° the action begins, and at 950° 10} per
cent. of the carbonic oxide is oxidised to carbonic acid.*

Traube’s Hypothesis.

That a trace of peroxide is found on the sides of a moistened jar
held over a lighted jet of carbonic oxide is put forward by Traube as

* Under other conditions the equilibrium may be very different. Thus Maquenne
(Compt. rend., 96, 63) found that when carbonic oxide and steam were submitted
to the silent electric discharge at a low pressure, 96 to 97 per cent. of the carbonic
oxide was turned into carbonic acid.

CARBONIC OXIDE AND HYDROGEN. 107

evidence that the burning of carbonic oxide is brought about by the
formation and decomposition of peroxide of hydrogen according to
the two equations—

(1.) co + H,0 + O, — Co, + H,0,,
(2.) 018) + H,0, as CO, + H,0.

Now as Traube considers that peroxide is invariably formed when
hydrogen burns with oxygen, the fact of peroxide being produced
when carbonic oxide burns in the air might equally well have been
used by him as an argument that there is hydrogen uniting with
oxygen in the flame, and that, therefore, the steam had been decom-
posed by the carbonic oxide. But as Traube had (wrongly) concluded
that carbonic oxide does not decompose steam at a high temperature,
such an interpretation was not open to him.

The fact that hydrogen peroxide is found under the conditions of
Traube’s experiment is no proof that the carbonic oxide is oxidised
by peroxide of hydrogen; the evidence advanced for such a reaction
is no stronger than could probably be advanced for several other
reactions. For instance, formic acid is produced when induction
sparks are passed through carbonic oxide and steam ; formic acid is
oxidised to steam and carbonic acid by oxygen; and dry carbonic

oxide and oxygen explode in presence of formic acid. These are facts
which might be urged with at least equal plausibility to show that
carbonic oxide is oxidised by the alternate formation and decomposi-
tion of formic acid according to the equations—

(1.) CO + H,O = H-COOH,
(2.) 2H-COOH + 0, = 2CO, + 2H,0.

This interpretation, however, of the function of the steam is not
open to us, because it has been shown that formic acid is not produced
either by a very powerful spark or by a red-hot platinum wire in
carbonic oxide and steam, whereas both the powerful spark and the
red-hot wire cause the explosion of moist carbonic oxide and oxygen.

The formation of hydrogen peroxide under the conditions described
by Traube seems to be due to the heating of the water by the carbonic
oxide flame burning in air, as I shall show in Part IT.

Part I1.—Tue Action or Hyprocen on OXYGEN.

The hypothesis I have put forward as to the manner in which the
presence of steam determines the explosion of carbonic oxide and
oxygen has thus been verified to this extent—that some of the steam -
present would certainly be decomposed in the region of ignition, with

the formation of carbonic acid and free hydrogen. It remains to
12

108 DIXON: THE COMBUSTION OF

answer the question, Does this hydrogen unite directly with the
oxygen present to form steam? Traube, who finds that peroxide of
hydrogen is formed in considerable quantity when the steam from
a hydrogen flame is rapidly condensed, by making the lighted jet
dip into water, concludes that hydrogen and oxygen do not unite
directly to form steam, but that they either react directly to form
peroxide of hydrogen, or that hydrogen, oxygen, and steam react
together to form water and the peroxide, which is in either case
then decomposed’ by hydrogen :—

(1.) H, + H,0 + 0, — H,0' H;0;,

It is certainly true that hydrogen peroxide is formed in considerable
quantity when a lighted jet of hydrogen dips into pure water. I have
confirmed this observation of Traube’s several times. But it is also
equally true that'a lighted jet of pure cyanogen gas playing on the
surface of pure water produces a considerable quantity of hydrogen
peroxide. Now, I have found that cyanogen, carefully dried by
anhydrous phosphoric acid, burns in:air or oxygen, the presence of
steam being unnecessary for its combustion.. The formation of
hydrogen peroxide appears, therefore, to be due to the heating effect
of the cyanogen flame; and the same: explanation accounts for its
formation by a lighted jet of hydrogen or of carbonic oxide. The
following experiment shows that when a portion of pure water is
evaporated in air, hydrogen peroxide is found in the residue. 10 cc.
of pure water quite free from peroxide was heated in a porcelain
crucible by means of hot sand. No flame or burning body was near
the crucible. When 8 c.c. had evaporated the residue gave the re-
actions of hydrogen peroxide strongly marked. It would seem, then,
that the formation of hydrogen peroxide is due to a secondary reaction,
and no argument as to the union of oxygen and hydrogen can be
drawn from its presence in condensed steam.

I have found that a mixture of hydrogen and oxygen dried by
long contact with anhydrous phosphoric acid is exploded by the
passage of an electric spark between terminals either of platinum, or of
silver, or of aluminium. The platinum terminals had been previously
heated in oxygen in presence of phosphoric acid, the silver terminals
had been previously heated in hydrogen in presence of phosphoric
acid, and the aluminium terminals had been subjected to both pro-
cesses, for the purpose of preventing the formation of steam at the
passage of the spark by hydrogen or oxygen occluded in the metal.
A platinum coil, previously heated to whiteness in oxygen in
presence of phosphoric acid, when raised to a red heat, after the
addition of hydrogen, instantly exploded the mixture. Although

Oo ere rr OF SS FF FR YW

CoO @oarrmr™~ FF @O CO ®D &

CARBONIC OXIDE AND HYDROGEN. 109

these experiments do not show conclusively that hydrogen unites
directly with oxygen at a high temperature without ‘the intervention
of steam, I think they make it probable that both under the influence
of the spark, and in the presence of red-hot platinum, hydrogen and
oxygen do unite directly. Once the reactions begun, no other steam
but that being formed in the reaction is necessary to propagate the
explosion. There can be no doubt, I think, on this point. For (1)
when a mixture of hydrogen and oxygen has its pressure gradually
raised until a spark causes an explosion, the pressure at which it
inflames is the same whether the mixture be well dried or damp; and
(2) the explosion travels as fast in a well-dried mixture as in a damp
mixture. Again, the experiments of Berthelot and Vieille (Compt.
rend., 95, 151) have shown that the rate of the “explosive wave ” in
hydrogen and oxygen is identical with the mean velocity of translation
of the steam molecules formed in the.reaction at the maximum tem-
perature of the explosion. This result I have recently confirmed bya
concordant series of measurements (Brit. Assoc. Reports, 1885).

These experiments show that when hydrogen and oxygen are
exploded in a tube, each successive layer is brought to the ignition
point by the impact of steam molecules from the layer just barnt.
The rate of the explosive wave corresponds with the velocity of
steam molecules, and not with the velocity of hydrogen peroxide mole-
cules, or with a velocity intermediate between these.two. Hydrogen
appears, therefore, to unite directly with oxygen to form steam when
the temperature of the mixture is raised; at all-events the reaction
takes place in the presence of steam.

The hypothesis of alternate reduction and oxidation by which I
have explained the action of the steam on carbonic oxide and oxygen
has thus been verified to the further extent, that each part of the
process has been separately verified. 1t has been shown that, under
the conditions of the experiment, carbonic owide is oxidised by steam with
the liberation of hydrogen, and that hydrogen -wnites with oxygen to
re-form steam.

It yet remains to show that these ‘two readtions give the best
explanation of the phenomena exhibited in the incomplete combustion
of mixtures of carbonic oxide and hydrogen.

The Incomplete Combustion of Carbonic Oxide and Hydrogen.

The experiments of Horstmann (Annalen, 190, 228; Ber., 10, 1626 ;
12, 64) and myself (Phil. Trans., 1884, Part II) on the incomplete
combustion of carbonic oxide and hydrogen have, I think, shown
conclusively that the final division of the oxygen between the two
combustibles depends on an equilibrium being established between the.

110 DIXON: THE COMBUSTION OF

two opposite chemical changes—the oxidation of carbonic oxide by
steam and the reduction of carbonic acid by hydrogen. In this paper,
a similar equilibrium has been shown to be established when mixtures
either of carbonic oxide and steam, or carbonic acid and hydrogen, are
heated by a platinum wire.

Though the establishment of these reactions removes the chief
argument for Traube’s hypothesis, it might yet be urged that the union
of oxygen with carbonic oxide and hydrogen takes place by means of
hydrogen peroxide, and that the reactions determining the above
equilibrium in an explosion are subsequent to the oxidation of the
carbonic oxide and hydrogen by the peroxide. The phenomena
observed in the incomplete combustion of carbonic oxide and hydrogen
are not in accordance with this view.

When a mixture of carbonic oxide and hydrogen is exploded with
insufficient oxygen for complete combustion, at a temperature at which
no condensation of steam can take place during the reaction, and at a
pressure greater than the critical pressure, the product of the carbonic
oxide and steam molecules remaining bears a constant ratio to the
product of the carbonic acid and hydrogen molecules remaining, what-
ever the quanities of carbonic oxide and hydrogen taken to start
with, and whatever the quantity of oxygen employed, so long as the
oxygen taken is less than half the hydrogen :-—

CO x H,O -_—
CO,x H

When the oxygen employed is more than. half the hydrogen, then
the ratio between. the two products (or the coefficient of affinity) is
diminished.

My experiments: have shown: that when a certain percentage of
oxygen is used, the proportion.of the hydrogen to the carbonic oxide
may be continually reduced without altering the coefficient of affinity,
until the hydrogen is reduced to twice the oxygen; at that point the
coefficient drops, and slowly diminishes as the proportion of hydrogen
is further reduced.. When a: different percentage of oxygen is em-
ployed, the same phenomena occur; the coefficient is coristant until
the hydrogen is reduced to twice the oxygen, and then the coefficient
drops as before. By varying the percentage of oxygen used the
lowering of. the coefficient may be made to‘occur with very different
proportions of hydrogen in the mixture; and since in all cases the
lowering occurs where the hydrogen and oxygen are present in the
proportion of 2 to 1, the one fact is shown to be necessarily con-
nected with the other.

Now, this alteration of the coefficient, when the oxygen is just equal
to half the hydrogen, follows as a natural consequence from the

CARBONIC OXIDE AND HYDROGEN. 111

hypothesis that the carbonic oxide is oxidised by the alternate reduc-
tion and re-formation of steam; it receives no explanation on the
hypothesis that carbonic oxide is oxidised by the alternate formation
and reduction of peroxide of hydrogen.

The presence of an inert gas, such as nitrogen, among the combus-
tible gases is found to favour the formation of carbonic acid instead of
steam, and therefore to lower the coefficient. If the carbonic oxide
receives its oxygen only from the steam formed, it follows that when
the oxygen taken is more than half the hydrogen, there will be
present during part of the reaction an excess of oxygen chemically
indifferent to the three other gases present, viz., carbonic oxide,
carbonic acid, and steam, and it will only be capable of reacting as the
steam is gradually decomposed by the carbonic oxide. No doubt the
molecules of steam, as they are first formed, begin to react on the
carbonic oxide, but the formation of steam in an explosion is much
faster than the formation of carbonic acid, so that during a period
which may be a considerable fraction of the whole time during which
the gases are reacting, the excess of oxygen is present as an inert gas,
diminishing the intensity of the reaction, like nitrogen, and so favour-
ing the formation of carbonic acid. We should expect then to find a
lowering of the coefficient to occur whenever the proportion of oxygen
approaches to half the hydrogen.

Now, what should be expected if peroxide of hydrogen were the
carrier of oxygen? If we suppose with Traube that the oxygen
unites simultaneously with hydrogen and steam, and. with carbonic
oxide and steam, forming steam, carbonic acid, and peroxide of
hydrogen :—

(a.) H, + H,0 + O, — H,0°+ H.0,,
(a’) co + H,0 + O, — CO, + H,0,,

and that the peroxide is reduced both by the earbonic oxide and
hydrogen in excess :—

(8.) H, + H,O, = H,O + H,0;
(p’) CO+ H,0, = CO, + H,0,

then, even if the formation of steam is considerably faster than that
of carbonic acid, there will be no inert oxygen present during the
reaction, for the oxygen will always be in the presence of carbonic
oxide and steam, with which it can combine directly. No lowering of
the coefficient ought, therefore, to occur. The same argument tells
against the “ contact theory ” of the action of steam in determining
the explosion of carbonic oxide and oxygen. If steam acted by its
mere presence without entering into chemical reactions with the
carbonic oxide and oxygen, there is no apparent reason for a sudden

112 ARMSTRONG: THE THEORY OF THE INTERACTION

change in the coefficient, for steam is present in all cases, and some-
times in larger proportion with the lowered coefficient than when
the coefficient is normal.

On the other hand, accepting Traube’s alternative view that
the oxygen unites directly with hydrogen to form peroxide of
hydrogen, then, if this reaction is the first to take place, we
might expect to find, in the absence of steam, an excess of oxygen
inert to the remaining gases, viz., carbonic oxide and hydrogen
peroxide, and only gradually entering into combination as steam
was liberated from the peroxide by the action of the carbonic
oxide. But in this case the lowering of the coefficient ought to occur
when the volume of oxygen is equal to the volume of hydrogen taken,
and not when it is half the hydrogen.

Post ScriptumM.—Professor Armstrong’s ingenious suggestion that,
in a mixture of carbonic oxide and oxygen, the carbonic oxide is
oxidised, and the oxygen is hydrogenised simultaneously by the steam
present is not, I think, opposed to any of the observed facts. It
would account for the lowering of the “coefficient of affinity,” when
the oxygen is more than half the hydrogen in the incomplete combus-
tion of a mixture of carbonic oxide and hydrogen, if we suppose that

the oxygen molecules are unable to react except when they meet with
both carbonic oxide and steam molecules simultaneously. Professor
Armstrong’s explanation involves the simultaneous occurrence of the
two reactions which I consider occur successively.

XII.—The theory of the interaction of carbon monoxide, water and
oxygen gases: a note on Mr. H. B. Dizon’s paper on the action of
carbonic owide on steam.

By Henry E. Armstrone.

Mr. H. B. Drxoy, in his paper on conditions of chemical change in
gases: hydrogen, carbonic oxide, and oxygen (Phil. Trans., 1884),
explains the action of steam in determining the union of carbon
monoxide with oxygen in the following words (p. 641) :—

*“When carbonic oxide and oxygen are exploded in a eudiometer,
the heat of the spark causes the carbonic oxide in its immediate
neighbourhood to decompose the steam . . . and the hydrogen
liberated by this reaction unites with the oxygen to re-form

OF CARBONIC MONOXIDE, WATER AND OXYGEN GASES. 113

steam. The steam so formed reacts with more carbonic oxide,
and so the alternate changes go on until all the carbonie oxide is
oxidised, according to the two equations: (1) CO + H,O =CO,
+ H,; (2) 2H, + O, = 2H,0.”

His account of the changes during explosion in a mixture of
carbonic oxide with hydrogen and oxygen is worded thus (p. 660) :—

“The explosion starts with the direct union of oxygen and
hydrogen, the change proceeding until all the oxygen is con-
verted into steam; as soon as steam is formed by this union it
begins to oxidise the carbonic oxide; the carbonic acid so pro-
duced is in turn reduced to carbonic oxide by free hydrogen.”

In the foregoing paper he has maintained these views in opposition
to Traube, and he has also referred to my interpretation of the nature
of the influence of water in inducing oxidation of the carbon mon-
oxide, contending that the lowering of the “coefficient of affinity ”
which is observed if oxygen be present in an explosive mixture of
H,, O. and CO to the extent of more than half the volume of the
hydrogen is an argument both against Traube and myself.

In my address to the Chemical Section of the British Association at
Aberdeen, speaking of Mr. Dixon’s discovery of the influence of water
on the explosion of a mixture of CO and O,, I said: “It appears to
me that the water may exercise the same kind of action as it (or
rather dilute sulphuric acid) exercises in a Grove’s gas battery, and
that its hydrogen does not become free in any ordinary sense.”

In a Grove’s gas battery we have dilute sulphuric acid in contact,
on the one hand, with hydrogen, and on the other with oxygen; it
may be supposed that before the circuit is closed the system is as
represented in (1) and that afterwards it is as represented in (2).

O|H,SO,|H, (2, OHs|SOuth

(1.) '
O H,S0O,'H, OH, SO,H,

The state before and after explosion of a mixture of CO, O, and
H,0 may be similarly represented, thus :—

O|H.0|CO OH,|OCO
OlH,OICO © OH,'0CO

Both Mr. Dixon and I regard the water as the interagent; but
while he considers that its oxygen becomes affixed to the CO, and that
its hydrogen in consequence becomes free, I regard the oxidation of
the carbon monoxide by the oxygen of the water as dependent upon
the simultaneous oxidation of the hydrogen of the water by the free
oxygen. (It is, however, I think, conceivable that the influence

114 GRIFFITHS ON THE USE OF

exerted by the water may be of a more purely “ mechanical ” order, viz.,
that it does exert a mere contact action, serving to bring together the
carbon monoxide and oxygen which are straining at each other.)

I do not see that any decision between our rival hypotheses can be
based upon the observation that a change in the ratio of CO, x H,
to CO x H,O—Mr. Dixon’s coefficient of affinity—takes place when
the oxygen is more than twice the volume of the hydrogen, as the
operations are quantitatively identical in the two cases; especially as
the coefficient of affinity only has a constant (maximum) value at
pressures above that at which its value is independent of the length of
the column of gas exploded, which varies for different mixtures.

One of Mr. Dixon’s general conclusions (No. 6, p. 675 of his

memoir) is that—

“The presence of an inert gas, such as nitrogen, by diminishing
the intensity of the reaction, favours the formation of carbonic
acid in preference to steam. When the hydrogen is less than
double the oxygen, the excess of oxygen cannot react with any
of the three other gases present—carbonic oxide, carbonic acid,
and steam—but has to wait until an equal volume of steam is
reduced to hydrogen by the carbonic oxide. The excess of inert
oxygen has the same effect as the inert nitrogen in favouring the
formation of carbonic acid.”

It may well be, however, that although the end-effect of the oxygen
is the same as that of the nitrogen, it is produced in another way;
and Mr. Dixon’s conclusion that oxygen is inert towards carbon
monoxide, even in presence of water-gas, is entirely based on his mode
of interpreting the character of the change. It appears to me that
he overrates the importance of the reaction between carbon monoxide
and steam: a series of experiments on the explosion of gaseous mix-
tures of carbon monoxide, water and oxygen would perhaps afford
information of value in regard to this point.

XIII.—On the Use of Ferrous Sulphate in Agriculture.

By A. B. Grirrirus, F.R.S.E., Lecturer on Chemistry, Technical
School, Manchester, &c.

In continuation of my experiments on the use of ferrous sulphate as
a manure, I have this year observed its effect on wheat crops with the
following results :—

FERROUS SULPHATE IN AGRICULTURE,

Conditions of Growth of the Crops.

. Two plots of land (exposed to the same rainfall and sunshine)
were chosen, both plots being well-drained land.

. The soils were of the same chemical and mechanical composi-
tion.

. One of the plots was manured with } ewt. of ordinary com-
mercial iron sulphate to the acre.

. Each plot of land was of the same size, and the same quantity
of seed was applied to each. The quality of the seed was the
same, being the produce of my last year’s crop.

At the end of the season, the following weights were obtained :—

Wheat Crop.

Plot of land manured with Plot of land, normal (i.e.,
iron sulphate. without iron sulphate).

(1.) Weight when |(2.) Weight{(1.) Weight when |(2.) Weight
gathered. when dry. gathered. when dry.

Total weight of crop
(grain + straw)... 6425 lbs. 5304 Ibs. 6460 lbs. 5316 lbs.

The crop of wheat grown by the aid of iron sulphate yielded
324 bushels of grain; and the crop grown without iron sulphate
yielded 30 bushels of grain. From the above results (although there
is an increase of 24 bushels of grain produced by the crop on the
iron-manured land), I am rather inclined to the opinion I gave in a
former paper (Chem. Soc. J., Trans., 1885, 46--55) that an iron
manure is not as valuable as a plant-food for cereals as for root and
leguminous crops.

On submitting the ashes of the plants from each plot of land to
careful analysis, I obtained results which do not differ materially
from those of last year.

Table I is an analysis of the ash of the entire wheat plants—

GRIFFITHS ON THE USE OF

Taste I.—The Entire Wheat Plants.

Grown with iron Grown without iron
manure.

Tron owide, FegO3....00+0ceceee

UN, Ma oc 0k co cccessceceoses
Boda, Nas .cccccceccccccccces
BE, GOED 060 csce cccesccacecees
Magnesia, MgO ......eccescesee
RN, EME. 6:0.00 00.00.00 0056 005606
Phosphoric oxide, PyOs 0.000000.
Sulphuric 7 _ sececececece
Chlorine . oeeewe

ior) —
tH C1 Co to bo

~
= one

The ash of the leaves gave the following percentage of ferric
oxide :—

Taste Il.—The Leaves of the Wheat Plants.

Grown with iron Grown without iron
manure. manure.

Iron oxide, Poe nvonsssnnseson 3°92 | 1°54

From the above analysis the percentage of iron oxide in the plants
grown with iron sulphate was about half a per cent. over those grown
on the normal plot of land. The percentage of ferric oxide in the
leaves of the crops is larger in those grown on the iron-manure land
by 2} per cent.

Besides my own experiments on cereal crops, Mr. George W.
Edgson, of Etton, near Peterborough (an agriculturist of great ex-
perience), has applied iron sulphate to several acres of wheat-crops
this year (1885). He tells me that this wheat-crop was better than
any he has had for very many years, and he “never saw such clean
straw”? in his life, being perfectly free from wheat mildew.

I had the pleasure of seeing these crops last summer, and may
remark that they looked the healthiest and finest in the neighbour-
hood. Ihave made analyses of the ashes from Mr. Edgson’s wheat-
crops grown with iron sulphate. The following is the percentage of
ferric and phosphoric oxides contained in the ashes :—

FERROUS SULPHATE IN AGRICULTURE,

Taste Ill.—The Entire Plant (Mr. Edgson’s Crop).

3°45
5°88

Tron oxide, FegO3 0.00 ceceseeees
Phosphoric oride, PgOg ..+ +0005:

Taste [V.—The Ash of the Leaves.

Tron oxide, Fe,O
Phosphoric oxide, Pz05.. ++ ++400.

From the above analyses, it is evident that there is a greater increase
of ferric oxide than is normally found in the ashes of wheat-crops,
and the phosphoric oxide is also increased. It appears that in this
case the iron manure has been benefivial—increasing the growth and
yield of the wheat-crop. Yet, as far as my own experiments with
cereal crops go, I have seen no: remarkable increase due to an iron
manure. These facts evidently show that some soils are deficient in
iron, or, to use the words of M. Ville,* “it may be that more efficacious
compounds of iron and manganese exist than those which the soil
contains naturally, and whose presence in manures would be followed
by an increased yield.”

Grass Land.

Mr. Edgson has also used iron sulphate to grass land. A large
plot of grass land of his was “infested” with the ordinary moss,
and was extremely poor in quality and quantity. After having applied
a top-dressing of iron sulphate to this land, the grass turned black
after the first rainfall, but in a fortnight became a bright-green
colour, and perfectly healthy, and the moss was destroyed. The
grass grew well throughout the summer. The bright-green ap-
pearance never altered in the least, although the summer was rather
a dry one, and other plots of grass land which had received occa-
sional dressings of nitrate of soda were “parched up,” more or less,
for want of rain.

I have submitted the ashes of the grasses and mosses to analysis,
with the following results :—

* “ On Artificial Manures,” Mr. Crookes’ Translation, 2nd Edition, p. 39.

GRIFFITHS ON THE USE OF

TaBLE V.—Ash of Mosses.

Ash of plants before | Ash of plants after
addition of FeSO, addition of FeSO,
to the land. to the land.

Iron oxide, Fe,O3.... 6 “62 11 ‘56
Alumina, Al Og... cc cccccccccecs 10°79 10°
2°48 3°
3°40 2°
11°28 10°
Magnesia, MgO.. co seecee 3°23 3°
Silica, is ce ccccccee 45°10 41°
Phosphoric oxide, PO; . 3°73 4°
Sulphuric popes ns cee cecccess ee 12°74 11°
Chlorine .. ce cccccccces 0°61 0°

99 ‘98

Taste VI.—Ash of Grasses.

Grass from land not Grass from land
manured with iron manured with iron
sulphate. sulphate.

2°46
23°12
6°10
14°26
4°98
29°61
7°21
7°43
4°82

Tron owide, Pegg oe ..cecececseeee
Potash, BO. oo 0000 cecess cccccese
Boda, NagQ...cccccccsccccesccece
Tne, Oa ccc ccccccccccccccccce
Magnesia, MgO........seeeeeeees
Bilica, SiO,.. 2... 2 cecccccecccces
Phosphoric oxide, PoVUs «++ ++00+++
Suiphuric oxide, SO, .......+..++-
Chlorine ....csccccccccecscecsees

~
CAanko

= > woo

99 -99

It will be seen from the above analyses (Table VI) that grass
manured with iron sulphate gives larger amounts of ferric and phos-
phoric oxides in the ash than the grass before the addition of iron
sulphate—hence the healthy appearance (during a rather dry summer)
may be due to this circumstance ; an iron manure being beneficial for
the growth of grass. The analysis of the ashes of the moss-plants,
after the addition of iron sulphate, shows that the percentage of
ferric oxide is 11°56. In a former paper (Chem. News, 50, 193,
“ Physiological Experiments with Iron Sulphate on Plants ”), I found
that all the plants I had examined died when they had absorbed iron
salts to the extent of yielding 10 per cent. of ferric oxide after incine-
ration. Hence the reason the mosses were destroyed (by the iron

FERROUS SULPHATE IN AGRICULTURE. 119

sulphate), being plants naturally requiring a large amount of
moisture.

Antiseptic Properties of Iron Sulphate and its Action on Vegetable
Parasites.

I have already published a paper on the antiseptic properties of
iron sulphate (Chem. News, 49, 279; also Ohem. Soc. J., Abstr.,
1884, 1070) in connection with micro-organisms and the Peronospora
infestans (potato disease). An aqueous solution containing as little
as 0'1 gram of iron sulphate in 100 grams of water has the power of
destroying the Peronospora infestans. After a careful microscopical
study, I have found that iron sulphate attacks the cellulose walls of
the hyphew of peronospora, perforating them. That is, it may act
chemically upon the cellulose—or the form of cellulose making up the
external walls of this lowly organism.

I have also made a careful microscopical study of the action of iron
sulphate on “‘ wheat-mildew” in both stages of its life-history. First
with the red spores (“rust”) of the barberry fungus, which produce
hyphe in the cells of the stem or leaf of a wheat-plant or similar
cereals; and also with the two-celled spores of the wheat-mildew
which germinate on the leaf of the barberry.

On placing the spores of “rust” and also the two-celled spores of
wheat-mildew on slides under the microscope, mounted in a drop of
water, then running in between the cover-slips the above solution of
iron sulphate, the spores are completely destroyed. Not only the
spores, but the hyphw which they give rise to, are also destroyed by
the same solution. Hence from this experiment it will be seen that
iron sulphate is a useful agent, besides being a direct plant-food, for
it has a remarkable action on parasitic life.

Turning to the more practical side of this investigation, I should
recommend the use of a weak solution of iron sulphate to water
ordinary farmyard manure before it is applied to the land ; for farm-
yard manure may contain spores from the wheat-straw forming the
litter in the farmyard. These and other spores hibernate until the
spring (farmyard manure being a medium that rather favours this
dormant state than otherwise), and are then ready to live their life-
history again.

Potato Crops.

This year I have grown potato crops with iron sulphate along with
other artificial manures, so as to make a comparison between potato
crops growing with the ordinary artificial manures used for potatoes
with and without the addition of iron sulphate. Three plots of well-
drained land were chosen, and the soil was of good quality. All were
exposed to the same climatic influences (i.e., rainfall and sunshine).

120 GRIFFITHS ON THE USE OF

The first plot of land (A) was left normal—not treated with any
manure. The second plot of land (B) was manured with—
1 cwt. kainite,
1 ewt. nitrate soda,
4 cwt. iron sulphate,
2 cwt. superphosphate of lime,
per acre. The third plot (C) was manured with the above consti-
tuents (per acre) minus the 4 cwt. of iron sulphate.
At the end of the season the following were the results obtained
(7 cwt. of tubers were sown on each plot of land) :—

TaBLe VII.—Potato Crops.

A. B. C.

2125 ,, 6231 ,, 4917 ,,

18940 lbs.

|
6720 Ibs. 19040 Ibs. | 14028 Ibs.
Total weight of crop.| 8845 Ibs. 25271 Ibs. |

From the above, the crop grown with the addition of iron sulphate
to the ordinary artificial manure, yielded 8} tons of potatoes (tubers).

The crop grown with the artificial manure without iron sulphate
yielded only 64 tons of tubers; while the normal plot of land only
gave 3 tons. Hence an iron manure is beneficial for the growth of
potato crops—largely increasing the yield and also, I may say in
passing, the quality.

I have submitted the ashes of the potato crops to analysis, with
the following results :—

Taste VIII.—Ash of Potato Tubers.

A. B. C.

Grown on plot
manured with

Normal plot. FeSO, and
artificial manure.

Grown on plot
with only
artificial manure.

5°42
53 69
1°20
3°19
7°23
5°21
16°24
4°00
3°81

Tron oxide, FezO3 .......
Potash, _ sede esesve
— Na,O .

nr
or

Silica, 50, coeee
Phosphoric enide, P,0,..
Sulphuric nena ~_*
Chlorine . “ one

as'oeowcu

SESSeSrss
“J Ors? ® e DO «7

i
Readrmmeood

~
Neeeannoes

99 -99

BS] co
&
ies
S

> = Pw a. Fi ted bee PO het i

FERROUS SULPHATE IN AGRICULTURE.

TaBLe [X.-—Ash of Potato (Haulm),

Grown on plot
Normal plot. mn | with only
eutitietes smeee, | artificial manure.
|

A. | 0.
| Grown on plot

Tron oxide, Fe,O3 .....+. ; 4°23
Potash, K.0.. .. secccess “98 27 86
Soda, Na,O 5°52 12°10
De EEE nina sk bene nk « . 14°92
Magnesia, MgO .. one “¢ 7°63
Silica, SiO... 3° 4°81
Phosphoric oxide, P, O;. "8: 9°92
Sulphuric oxide, #0. , ; 5°90
Chlorine .......... anes 2°% 12°62

99°99

It will be seen from the analyses of the ashes of the plants that
the ferric oxide is increased in the ashes of the crops grown with an
iron manure, and also the phosphoric oxide increases as the iron

increases. The potato crop grown with the artificial manure, plus
}ewt. of iron sulphate, gave a much larger yield of produce than
either of the other two crops.

Retentive Properties of Iron Sulphate for Ammonia and Phosphoric
Acid.

From my previous papers on this subject, and also the present one,
it will be observed that all those crops grown with iron sulphate gave
an increase of phosphoric acid, P,O;, in their ashes over those not so
manured. Therefore I am inclined to think, from these practical
results, that iron sulphate (or the iron manure) has a retentive pro-
perty for phosphoric acid contained in soils. Ferrous sulphate has
also a retentive property for ammonia contained in soils. This I have
proved experimentally, for by preparing a thick layer of soil (placed
ina wide glass tube about one yard long) containing iron sulphate,
and then allowing a solution of ammonium sulphate to percolate
through this layer, a large amount was absorbed—the liquid which
came through the soil containing much less ammonium sulphate than
the amount in the original solution. Again, when the same quantity
of soil (minus the iron sulphate) was treated exactly in the same way,
4 much larger amount of ammonium sulphate was found in the
percolated liquor.

VOL. XLIX. k

122 THE USE OF FERROUS SULPHATE IN AGRICULTURE.

Duration of Iron Sulphate in the Soil.

It was at the suggestion of Mr. Bischof, F.I.C., that I shouid
ascertain how long after manuring a soil with ferrous sulphate any
ferrous sulphate can be detected in the soil. I have found that after
manuring a soil with ferrous sulphate it could be distinctly detected

in the soil after the lapse of six weeks.

Method of Applying Iron Sulphate to the Land.

The question ‘‘ Which is the best way to apply iron sulphate to the
land ?” has been repeatedly asked by numerons inquirers.

I. I have found it best to apply the iron sulphate as a top-dressing
to the land after the crops have appeared above the ground. The
sulphate is capable of withstanding (during a comparatively dry
season) beyond a mere superficial oxidation, for some weeks on the
surface of the soil—it being gradually dissolved and passing into the
soil.

II. The amount of iron sulphate I have used in all my experiments
is 4 cwt. to the acre.

Sir John B. Lawes, F.R.S., kindly writes me, that he has found

that 14 ewt. of iron sulphate per acre “ was rather too much, although
it did not kill the plants.”

Experiments with Iron Sulphate on Rose-trees.

Last spring I treated several young rose-trees (climbing, about
2 feet in height from the soil) with iron sulphate. By the end of
July they had aattined the height of 16 to 20 feet—increasing almost
in the same proportion laterally. The rose-trees produced numerous
roses of the most perfect form and colour. All new buds produced
by these trees resisted the attack of the “green fly.” In former
years young rose-trees on the same land grew very slowly—some never
flowering at all. From these experiments, and also others performed
by several friends, I am able to say that iron sulphate is a good
manure for rose-trees.

123

XIV.—On Multiple Sulphates.

By Emity Aston and Spencer Umrrevitte Pickertne, M.A. Oxon.,
Professor of Chemistry at Bedford College.

THE double sulphates of the type M”SO,,M',SO,,6H,0, in which M”
represents any metal belonging to the magnesium or copper class, are
stated to be able to combine with each other, molecule with molecule,
forming multiple sulphates such as

(CuSO,,K,S0,,6H,0,MgS0,,K,S0,,6H,0),
(CuSO,,K,S0,,6H.0,MgS0,, K,S0,,6H,0, FeSO,,K,S0,,6H.0),
&c., for a description of which we are indebted to Vohl (Annalen, 94,
57), as many as 42 of these “ double double” sulphates having been
described by him (half of them containing potassium sulphate, and
half ammonium sulphate), as well as two “triple double” sulphates,
and one “quadruple double” sulphate. They may be regarded as
being derived from the double magnesian sulphates such as
CuMg(SO,).,14H,O, by the substitution of 2 mols. of potassium
or ammonium sulphate for 2 mols. of water, just as the double salts
such as MgSO,,K,SO,,6H.,O may be regarded as derived from the
single sulphate MgSO,,7H,O by the replacement of one of the water
molecules in it. Of the double magnesian sulphates, Vohl described

two preparations, as well as some double and triple alums.

Being desirous of investigating these complex double salts, with a
view to their bearing on the constitution of molecular compounds in
general, we attempted to prepare some of them by the simple method
employed by Vohl, namely, mixing solutions of the constituent sul-
phates in the requisite proportions, and leaving the mixture to
evaporate spontaneously. The composition of the salt thus formed,
however, by no means corresponded to the theoretical numbers, and
we were consequently led to undertake an investigation of these sub-
stances; this has led to the unexpected conclusion that none of them
can well be regarded as definite chemical substances. It is certainly
with some diffidence that we venture to throw doubt on the sur-
prisingly accurate analyses of 50 different salts published by Vobl,
and we are even unable to explain his results by the fact that he
invariably mixed the constituent sulphates in the theoretical propor-
tions, and did not vary the conditions of his experiments, for the very
methods employed by him have failed to yield us definite compounds.
We can only publish our own results without venturing any explana-
tion of his.*

* With reference to the double salts, see also Hannay, Chem. Soc. J., 1879,
Trans., 456 ; Pickering, Chem. News, 52, 251.

K 2

Dime eiG Bae 28 tg

124 ASTON AND PICKERING: ON MULTIPLE SULPHATES.

The “double double” salts were the first objects of our inquiries,
and three methods of preparing them were adopted, varying in each
case the proportions in which the constituents were mixed.

The first of these methods consisted in mixing cold saturated solu-
tions of the constituent sulphates, and examining the crystals which
separated on standing, without allowing any evaporation or change
vf temperature to occur. The second method was that of cooling
hot saturated solutions; and the third, allowing the solutions to
evaporate spontaneously.

The first salt chosen was the potassium copper magnesium sulphate,
of which the theoretical formula is

(CuSO,,K.S0,,6H,O,MgS0O,, K,S0,,6H,0) .

The method employed in ascertaining the composition of the salt con-
sisted in determining the amount of copper which it contained, this
being the only method which could be employed satisfactorily, since
the determination of magnesium is most uncertain, and a wide range
in the composition of the salt would affect the sulphur or water
percentage to a small extent only.

The results of two series of experiments are given in Table I: the
wide difference between them is due to the fact that they were not
performed at the same temperature, but in themselves they are con-
cordant, and show conclusively that the composition of the crystals
varies continuously with the proportion of the constituents taken, and
that in no case did any of them agree with the theoretical compound,
which should contain 9°375 per cent. of copper oxide; the nearest
approach to this was obtained in Experiment 5, where the magnesium
taken was over two and a half times as much as the copper. In this
and every other case, the potassium sulphate added was sufficient to
combine with the whole of the other sulphates present: the actual
amount of potassium sulphate taken being 23°7 grams. The colour of
the crystals, as well as the amount obtained, varied with the propor-
tions taken.

Table II contains the analyses of salts obtained by cooling hot
saturated solutions: in one case only did the crystals correspond in
composition to a definite compound, but the other experiments show
that this was fortuitous, and that a variation in the proportion of
the salts taken produces a variation in the product, just as in the case
of cold solutions. A comparison of the results given in this table
with those in the previous one shows that temperature has a con-
siderable effect on the nature of the product.

The figures in Table III show that the products obtained by spon-
taneous evaporation are not more definite than those obtained by the

other methods.

ASTON AND PICKERING: ON MULTIPLE SULPHATES. 125

TaBLE I.—Potassium, Copper, and Magnesium Sulphate Solutions
miced.

Crystals obtained

Solution taken

contained Cu : Mg. Contained

Weighed.

| Percentage CuO.

15°0 16°518
10 °6 14° 863
2°9 13 °050

11°205
9°614
5°891

Taste IIl.—Hot Solutions of Potassiwm, Copper, and Magnesium
Sulphate cooled.

Crystals contained

Solution taken
contained Cu ; Mg.

Percentage CuO. | Cu : Mg.

14°709
12°478
9 °433

TaBLe III.—Potassium, Copper, and Magnesium Sulphate Solutions
evaporated over Sulphuric Acid.

Crystals contained
Solution taken ; eres ze
contained Cu: Mg. |
| Percentage CuO. | Cu : Mg.

Po. Bee 13°958
eee 10 “965

The next salt which was investigated was the supposed copper
cobalt compound CuSO,,CoSO,2K,S0,,12H,0.

The results obtained on mixing cold saturated solutions of the con-
stituent sulphates as given in Table IV coincide perfectly with those

126 ASTON AND PICKERING: ON MULTIPLE SULPHATES,

yielded by the copper magnesium sulphate; the quantity of the
crystals, their cclour, and their composition varying regularly with
the proportions taken, and showing not the slightest indication of the
existence of a definite chemical compound. The variation in colour
was so distinct, ranging from a dirty blue in Experiment 12 to a
strong red in Experiment 19, that analyses were almost superfluous,
and only four out of the eight products were analysed.

Taste 1V.—Potassium, Copper, and Cobalt Sulphate Solutions mized.

Crystals obtained

Solution taken ;
contained Cu : Co. Contained.

Percentage CuO.

14°016

5
3
2
1°5
1
1
1
1

ON Re ee ee
Fannin qn- 17
CHWASHABOGD

The results obtained by cooling hot saturated solutions are given in
Table V, and those obtained by spontaneous evaporation in Table VI.
A regular variation in the colour and composition of the crystals is as
well defined here as in the other cases, The solutions mentioned in

Taste V.—Hot Solutions of Potassium, Copper, and Cobalt Sulphates
cooled.

Crystals contained
Solution taken
containea Cu : Co.

Percentage CuO.

9-874
Same in colour as No. 20.
Intermediate in colour between |
Nos. 20 and 23.
6° 169
Similar in colour to No. 23.
As red as No. 19.

* With deficit of potassium sulphate.

ASTON AND PICKERING: ON MULTIPLE SULPHATES. 127

Table IV were also allowed to evaporate spontaneously, and the
crystals thus obtained in every case resembled so nearly the first crops
as to be indistinguishable from them.

Taste VI.—Potassium, Copper, and Cobalt Sulphate Solutions
evaporated spontaneously.

Crystals contained

Solutions taken
contained Cu : Co.

Percentage CuO. Cu : Co.

re : 9°78 1°3:1
1 :1°

WT sees : 6°918 6

28 .... . All exactly the same in colour as those obtained on
29 .... mixing the saturated solutions in these proportions.

BO secs : .
31... See Table LV.

Table VII contains the analyses of some crystals obtained on eva-
porating solutions of potassium, cobalt, and magnesium sulphate ; these
also do not correspond to any precise formula, and become poorer in
cobalt as the solution yielding them contains less of that sulphate.

Taste VII.—Potassium, Cobalt, and Magnesium Sulphate Solutions
evaporated spontaneously.

Solution taken contained Pee ae ‘
Co : Mg. Crystals contained Co : Mg.

382 .... 1:1 first crop
33 2.0 »» second crop

* Determined by estimating the water and the sum of the magnesium and
potassium sulphates present.

Such being the results obtained with the ‘“‘ double double ” sulphates
of the supposed type [M’SO,,K,S0,,6H,0,M;’SO,,K,S0,,6H,0 ], it re-
mained to be ascertained whether the salts from which they were
said to be derived, {M'SO,,7H,O,M;'SO,,7H,0], were equally hypo-
thetical.

When saturated solutions of copper and magnesium sulphates are
mixed, no crystallisation takes place, consequently the mixtures of
these solutions had to be further evaporated. Table VIII contains
the results of evaporating them while hot, and allowing them to cool.

128 ASTON AND PICKERING: ON MULTIPLE SULPHATES.

Nos. 34 and 35 crystallised in the same form as copper sulphate,
and a determination of the water present showed that both the sul-
phates constituting them contained 5H,0, whereas in Nos. 36 and 37
they took the form of magnesium sulphate crystals, each containing
7H,0.

Taste VIII.—Hot Solutions of Copper and Magnesium Sulphates
cooled.

Crystals contained

Solution taken
contained Cu : Mg.

Percentage CuO. |

31°438 |
28°71
11-716 |
9°612 |

These sulphates show as little tendency to combine in definite pro-
portions as those containing potassium, the composition of the pro-
duct varying continuously with the proportion in which they are
taken: indeed, there even appears to be a marked disinclination
to crystallise in equal molecular proportions: it was noticed that the
solution containing 1 mol. of each of the metals required much more
evaporation than any of the others before crystallisation could be
induced, as if, while the opposing forces were equal, the sulphates
were uncertain whether to assume the form of the copper or of the
magnesium salt, and finally, when they did crystallise, the proportion
in which they did so was far from 1:1 mol. A similar tendency to
avoid a definite composition is observable also in the results obtained
by the spontaneous evaporation of these salt solutions as embodied in
Table IX. The water determinations are here given, but, owing to
the fact that a considerable alteration in the composition of the salt
would cause but a small alteration in the percentage of water present,
this does not give such a trustworthy method of calculating the
composition as the copper percentage does.

Experiments 47 and 48 were performed on very large quantities of
the salts, and at a different temperature from that employed in the
other experiments.

The results of this examination would prove that salts correspond-
ing to either of the formule

[(M"SO,,K,S0,,6H,0) (M{’SO,,K.S0,,6H,0) ]
or [(M"’SO,,7H,0)(M,’S0,,7H,0)] do not exist, and that similar

ASTON AND PICKERING: ON MULTIPLE SULPHATES. 129

TaBLeE [X.—Copper and Magnesium Sulphate Solutions evaporated
spontaneously.

Crystals contained

Solution taken
contained Cu : Mg.

Percentage CuO. | Cu: Mg. | Percentage H,O.

—- | 36-183
2nd crop | 30°87 (33°5: =
— 36 °328
: 47 °438
47°416
47 °755
48 ‘077
48 °332

48 °773

2nd crop | 11 °655
3rd crop | 10°890
y 11°726
2nd crop | 10°813
9-693 ,
; ‘608 9-649
2nd crop| 8°323

48°171

1:2 12°745
» 2nd crop | 416

mo | ~ nm @

ont

salt molecules, such as are here supposed to be united, have no
tendency to combine with each other, were it not for two facts, which
tend to show the contrary. When saturated solutions of potassiam
and copper sulphates are mixed, a certain amount of the double salt
separates; in the case of potassium and magnesium sulphates this is
not so (this vol., p. 16); but on mixing all three of these sul-
phates the copper salt carries down with it a considerable quantity
of the magnesium salt. The second fact is that, with the copper and
magnesium sulphates, either, when in excess, will induce the other to
crystallise out in a form, and with a proportion of water foreign to its
nature.

These are the only two pleas which we can put forward for regard-
ing either of these salts as chemical compounds ; and, even if they be
accepted, we can certainly not call them definite compounds, but
substances more nearly resembling the product of the crystallisation
of two isomorphous salts.

Note.—Since writing the above, my attention has been called to an
investigation of the potassium cobalt nickel sulphates by J. M.
Thomson (Brit. Assoc. Rep., 1877, 209), the results of which are
entirely in accordance with the experiments here given. Thomson
found that the successive crops of crystals deposited on evaporating a
saturated solution of the component salts at 80°, varied greatly in
composition ; in no case did they correspond, even approximately, to

130 TURNER: THE INFLUENCE OF SILICON

the formula given by Vohl, and they could only be represented by
a formula by assigning to it a degree of complexity comparable
with the errors of analysis. At the same time, there was evidence
that these crystals could not be regarded as mere mixtures, for the
dichroism exhibited by them showed that the composition of each was
uniform throughout.

XV.—The Influence of Silicon on the Properties of Cast Iron.
Part III.

By Tuomas Turner, Assoc. R.S.M., Demonstrator of Chemistry,
Mason College, Birmingham.

In the two former parts of this paper, an account has been given of a
series of experiments undertaken with the view of determining the
influence of silicon on cast iron. The results obtained were, in some
respects, contrary to what might have been anticipated from the
evidence on this subject which was previously in our possession. It

became of considerable interest therefore, to ascertain whether these
results were supported, or otherwise, by the work of other investi-
gators; and whether, on the large scale, the same effects could be
produced as had been noticed in these experiments condacted with
but a few pounds of metal in the laboratory.

The chemistry of iron manufacture, as befits its great practical im-
portance, has received much attention ; and in the list of workers in
this field may be found the names of many of our most illustrious
chemists and engineers. But it would appear that the importance of
combined chemical and mechanical testing has only of late years been
fully recognised. Thus in the earlier experiments on the strength of
cast iron, conducted upwards of half a century ago, the object ap-
peared to be simply the determination of the limits of strength, and
of the general average of a large number of specimens. ‘This class of
tests is doubtless of considerable value to the engineer, but to the
chemist it is almost worthless. An advance was made some 40
years ago in the very important experiments of Fairbairn and Hodg-
kinson, by whom a large number of hot and cold blast irons were
compared, and tests made upon iron of different grades from the same
locality. This work entailed years of skilful, patient labour, and the
results were of immense importance; but it is to be regretted that
they were not accompanied by chemical analyses. The iron produced
in any given locality often changes very much from time to time,

ON THE PROPERTIES OF CAST IRON. 131

varying even in the same furnace from day to day; and we have no
reason to suppose that the results obtained from an iron of certain
locality and grade in 1844 would be of any importance as a guide
to-day. If, however, analyses had been given, we should then have a
standard for comparison much more permanent and important than
that furnished by mere locality, and the value of such information to
the practical chemist could scarcely be over estimated.

A great advance was attempted about 30 years ago in the work
of the American Commissioners on the strength and other properties
of metals for cannon, whose report appeared in 1856. In this report,
chemical analyses accompanied the mechanical tests, and much benefit
was anticipated from this combination. The mechanical tests afford
considerable support to my own conclusions, especially in regard to
tensile strength and hardness. In the best specimens examined, the
observed tensile strength considerably exceeded the maximum found
in my experiments, and which has been objected to in some quarters
as being exceptionally high. The range of hardness also appears to
fairly correspond with my own observations. In the case of their
chemical analyses, however, the experimenters appear to have been
less fortunate. Three chemical reports were issued, and in the last
or “ Final Report on Chemical Analyses,” p. 394, we read :—

“Upon a comparison of this table with those heretofore presented
during the progress of our researches, it will be observed . . . that
the figures obtained by us for a few samples are discordant with
those in the former tables. We are not aware how these errors may
have occurred, but that they do exist is sufficient motive for our
rejection of all the results, for where a few are wrong, all may justly
be suspected.”

The following examples will sufficiently illustrate the character of
the discrepancy.

No. 28 C.F. 42-pr. No. 10 W.P.F. 32-pr.

Specific gravity, 7-220 Specific gravity, 7 *245
Tensile strength, 31,734 lbs. | Tensile strength, 31,734 lbs,

2nd Report, | 3rd Report, 3rd Report,
p- 385. p. 396. . . p. 396.

Allotropic carbon ...... 2°20 2°20
Combined carbon 1°70 1°70
ET snp esanhaseo 0°329 0°329
0-000 trace
0-033 -—-
10°31 0°436
1°67 3 °547

ae @ene
$3! ese

132 TURNER: THE INFLUENCE OF SILICON

Though these results have been largely quoted in certain text-books,
I have considered it best, under the circumstances, to imitate the
example of the authors themselves, and not to draw any conclusions
from unsatisfactory data.

In 1856 a most valuable series of experiments was commenced at
Woolwich, and the results obtained are published in the report, ‘‘Cast
Iron Experiments, 1858.” This report probably contains the most
complete information on British cast iron which we at present possess.
As comparatively little systematic work has been done in this direc-
tion since the report was issued, it is worthy of more than passing
attention; I propose therefore to examine the results in detail, in
order, if possible, to show the influence exerted by silicon on the
properties of the metal.

The chemical examination in these experiments was directed by
Sir Frederick Abel, almost the whole of the analyses being performed
by Mr. John Spiller; and on the most critical examination the work
appears to be in every way worthy of the department by which it
was conducted.

In this report, 70 different irons were examined, seven of which
were of foreign origin. In most cases, determinations of the im-
portant chemical constituents are given, while an elaborate series
of mechanical tests is also recorded. Analyses of ores and fluxes
were performed in many cases, but with these we are not now con-
cerned. For details as to the methods adopted, the shape and size of
the test pieces, and other similar information, the report itself must
be consulted, for in this place we shall be able only to deal with a
short summary of the results. It is to be regretted, however, that
the shape of the test pieces adopted for the determination of tensile
strength, was not suitable for the accurate measurement of elongation ;
this fact is noticed in the report itself (p. 9), and has prevented my
calculating the modulus of elasticity, which would otherwise have
been of great interest, on account of the exceptional values obtained
in some of my own experiments (comp. this Journal, Trans., 1885,
580 and 583).

In looking over the report and appendix, it will be found, that of
the 70 irons experimented upon, eight are mentioned as being “ too
hard to turn.” These are given in the following list (Table F, p. 133).

It will be noticed that in seven cases special hardness is accompanied
by a low percentage of silicon; while in the mottled iron, where the
silicon is present in sufficient quantity to produce very soft metal in
a fairly pure iron, there is upwards of 1 per cent. of phosphorus, and
this would quite account for its unusual hardness. This specimen is
also richer in sulphur than any other iron examined in these experi-
ments, this also would account for its greater hardness. These results

ON. THE PROPERTIES OF CAST IRON.

TasLe F.—Jron too Hard to Turn.

I. Britisu Irons.

Locality, &c. Silicon. Phosphorus.

Netherton, No. 5 strong forge........+.-+eeseeees ‘57 “29
Ulverstone. . POreTerererer erie “59 10
Wellingborough, } ‘No. 2 mottled . eo eeeecccevese ‘ll ‘07

II. Foreran Irons.

0g te De gal Ey eT eT ee ee ‘21 *53

ee. cae eee kabahina esaediy eek aan "44 *27

Elba. . PTET TTS TTT TST TOE TT TTT 71 ‘07

Swedish grey .. (Aki RKO nS ERAN Ee be Kae ESS "86 ‘10
Tuscan 88 ‘09

afford the strongest possible confirmation of the softening effect pro-
duced by a certain amount of silicon; an effect, so far as I am aware,
first pointed out in my own experiments. Hitherto it has been the
general custom to attribute an entirely opposite influence to this
element.

The seven foreign irons mentioned in the appendix form a very
interesting series, when arranged in order of silicon (Table G).

Taste G.—Foreign Irons (Arranged in Order of Silicon).

| Chemical analysis. |

Trans- | ,
No.| Locality, &e. verse | oe
strength. qualities.

1 | Nova Scotia white . | O° | 2 96 0°02) 1°5 Too hard
} | to turn.

2 | Indian charcoal .../| 0° 14 0° 44 3°38} — | 0°04 0°27 2 Too hard
to turn.
ne ae — | 0-7 2-0 0-03; 0-07 Too hard
to turn.
3 ~ 0°03) 0°10 7 Too hard
| to turn.

|
4 | Swedish Brey ...00| °° 17 0°86!

2-91 0°08 0-09 Too hard
to turn.

5 | Tuscan a 6°88

soft.

7 | New York, sterling} 0 os 1°33) 2°87 a 0°09 Good and
, soft.

|
6 Nova Scotia grey ..| 0° 25 es ha 3°29 0°01) 0°13 Good and

134 TURNER: THE INFLUENCE OF SILICON

This list contains all the foreign irons mentioned in the report, and
very well illustrates the effect of silicon on the tenacity and working
qualities of cast iron, entirely confirming the results of my own ex-
periments. It will be seen that the first five specimens are too low in
silicon, and are consequently inferior both in strength and working
qualities; while even the two last, which contain most silicon, and
are the best irons in the series, do not quite reach the best proportion
as deduced from my observations. The transverse test was the only
one which could be applied in the first five specimens, poor in silicon,
owing to their great hardness.

We will now take a general survey of the results given in the
report, as distinguished from the appendix, with the object of making
clear the influence of silicon on the mechanical properties of the
metal. This part includes 53 samples, two of which were too hard
to turn, and in which the mechanical tests were consequently incom-
plete. In one other case, the analysis is not given. Hence in each
series of tests we have 51 to 53 samples, and these are arranged in
the report in order of quality. It will be obvious that in such a case
the average order of quality of the whole series would be about 26.
Any specimen in the list above this number may be considered as of
more than ordinary quality, while those below 26 are of inferior quality.
For comparison I have selected three sets, each containing six speci-
mens, viz., the six lowest in silicon, the six best specimens, and the six
highest in silicon. The analyses are given side by side with the order
of quality in sp. gr., and tensile, transverse, torsional, and crushing
strengths ; for these latter a mean is obtained, and a general average is
deduced for each series (see Table H, p.135). The results are as fol-

lows :—
Average Si per § Average order
of quality.
Six specimens lowest in silicon “98: 16-0
Six best specimens 3 6°4

Six specimens highest in silicon .... 3° 42°3

These numbers show conclusively that the specimens Jowest in silicon
are not the best. It is true that they are above the average in order
of quality—we have seen the general average is about 26—but this is
accounted for by the fact that no white irons were examined, and so
the amount of silicon present is not much below the most suitable
proportion for general strength. The six best specimens contain, on
an average, 1°393 per cent. of silicon; and it will be remembered that
my own experiments showed the most suitable amount to be about
1-4 per cent., so that the agreement is remarkably good. In the six
specimens highest in silicon, the ill effects of too large a proportion is
very marked.

pp. 154-5.

Taste H.—From “ Cast Iron Experiments, 1858,”

Zz
)
=
—
ee
77)
-
oO
a
°
72)
|
e
fa
=
jo}
om}
os
q
8]
e
a
°

*2-4@ = Ayyend jo sopio oBv10a¥ 7

“IS “quedo Jed 1.1

Puy “4 “jue Jed 4Z.] 98 YONUI se pouTeyUCO 41 Be ‘popnioxe sf ,, Bid pousyFnoy ‘oaqsoippry_ ‘yUeG YINOG,, ‘UOoTIs Jo Jopso Uy ySeMO] XIs OY} BuTye, USyYM—"A70KT
*¢.2% = Aqyenb jo Jopio eBvi9vaAy “gF.g = UOoTIs “Juae Jod oBvlsay

Mow IG

€9.% | 90813 | *Suyjour yeiy ‘uo-F1g |" ““uopeay ‘sy M psojfoy

8t-€
it ¥
20- €
11-3
€1-%

** Bid Aspunoy Z “ON

** uosl-3id g ON
11-0 * Bid Aspunoy | “ON
10-0 Sid sapunoj ¢ “ON
8e-0 [*""** Bid Aapunoj Z “on

0€-0
40-1

* S81PNC ‘tH PLO

““TOABYOUY M ‘oz1;BUIeey
** ,o1gsa|ppIA ‘queg qynog

‘(uooyrg fo sapig ur) uooyrg ysaybrT yyim suauedg xg ‘TIT

“7.9 = Aqyenb jo sopso oBesaay “EGE.[ = UOdTISs *yue0 Jed oFe10Ay

*usmpods Ay[Ne 7,

*@AOQU PossEy

qnoggip A108
oe
“
Ti

#08 L
12 | % ¥
G zZ

I £

£ 9

vA G

ral

12-0 | 80-0
61-0 | O1-0
8F-0 | O1-0
1g-0 | $0-0
¥£-0 | 90-0
21-0 | ¢0-0

| 89-1
88- I
121
£8-0
0-1
92-1

‘dee uowlul0g
pes sseeereesee #+* DogigUOy
8" @B40g

9310} G pue F ‘ON

terres Sid Zuijjou Z “ON
severe Bid Zuljjoul g “ON

99.
€l-
98.
12.
¥-

***"9**"O IO M UOT UOARUOLTA
“***"""OC) UOT OL BPlBIM

sreverevees KQipnq ‘peey ye

seveeeees SYIOM UOL] UOJIOION
“ “ os

“* grrysi{qiog ‘Urey IS0 MK

‘(igyont) fo saps— Ur) sarsag ay} us suamwady 48aq xg “]T

“9

*uoujoods ANV 7,

pooy
4308 0q 0} poutfouy

*soraenb Zuryi0

61 | ee

0% | &%

“aysuaL | gre~

*£yyenb jo 1epigo

= Ayyenb jo Jopso oFvs9Ay «*ER6.0 = UOdTTIS “Jude 10d OFe10AV

t¥-0
ge. 0
62-0
Rg. 0
1¢-0
8e-0

$2SS2S

60-1

96-0
£8-0
18-0

oooooo

91-1
GO-1 | 1-1

pisencsonaas kai € ‘ON
988[q PlO9 ¢ “ON
seeeeeeees HOUT QNBUIRH
* Sid Zuynjeu Aspunog
03105 G pus F ‘ON
0310} 9 puv | ‘ON

¥6-0
99-0

c¢.0
12-0
€¢-0

*reeeee"* SHIOM UOJ] UOLIONION
"SIO M UOIT UOABUDRIG
~*£91png ‘sys0yy shir]
seers SOipng ‘peay eq |

** SHJOM UOI] UOLIONION

“"""Sy10OM UOL] UOABUERIg

*uoryd10seq

“sisA;eue [BorMaYyD

*porjddns woya Ag

9
g
¥
€
d
I

*(uovynrg fo 4eps—Q ur) UuooNIg) gsemorT yzIm suawmmedg wig “Tt

*Q-

PST

‘dd ..‘.g¢eat

‘squoultsedxgy UoAy 48¥BQO ,,

moig—FL ATAV I,

136 TURNER: THE INFLUENCE OF SILICON

But however conclusive such a general view as the above at first
appears, it might be supposed that a closer examination of the report
would not support the conclusions drawn. We shall, therefore,
proceed to consider the report in detail, and to examine the results
obtained in each specimen examined. Such a task is one of no small
difficulty, and a number of irregularities must be confidently antici-
pated. These are due in great part to the fact that we have to do
with at least six variables, namely, manganese, silicon, graphitic and
combined carbon, sulphur, and phosphorus. Each of these sub-
stances exerts a considerable influence on the character of the product,
and we have little knowledge of a definite character as to the in-
fluence of different proportions of these elements, either alone or in
presence of each other, on cast iron. Added to this, we have the
errors due to experimental observation, both chemical and mechanical,
which in several hundreds of analyses, and a much larger number of
mechanical tests, must be by no means inconsiderable. In many
cases also the differences in silicon are but small, sometimes less than
0'1 per cent., so that its effect is entirely overcome by variations in
other elements present. Lastly, it would appear that the analyses
were made on the samples as received, which was probably the best
way under the circumstances; but the result is, that they do not
accurately represent the composition of the actual pieces tested, owing
to slight alteration during melting; this introduces a considerable
disturbing influence when the composition of the metal is very nearly
the same in several specimens. Hence, if we are successful in show-
ing the general influence, and in accounting for most of the im-
portant exceptions, we shall accomplish all that can be reasonably
anticipated.

In order to eliminate, as far as possible, one source of variation, and
to reduce the number of specimens to be compared together at one
time, I have divided the results according to the amount of phos-
phorus present. It would be obviously unfair to compare together
two specimens differing from each other by less than 0:1 per cent. of
silicon, and by over | per cent. of phosphorus, as the influence of a
small variation of silicon would be entirely masked by the large
variation in phosphorus.

I. Non-phosphoric Pig-iron.

Beginning in order, we take first non-phosphoric pig-iron. In the
report we have nine specimens, each containing under 0°20 per cent.
of phosphorus; these are arranged, in order of silicon, in Table K.
It will be seen that all the specimens possessed good working quali-
ties, except the first, which was the lowest silicon and highest sulphur

:
3
m4
Q
>
S
Fy
a
~
<q
~
~
2
‘-
ty
~
ond
<
o
o
$
&
a
°
h
Q
™~
3
«
3
w
§
N
q
‘oe
‘
dQ
ry
®
-
A
5
=
by
9
Ld
by
s
z
a
ic)
r)
<q
iS

“anydyns 1943.4
“auydes3 ysoysiy

@ w

908 9q 04 paulfouy
4J08 9q 04 pourtfouy
oo

poop
qmowrp £104

poon

8887282 | 83

poor

poop

CwUnBDnwaovowonrooe

ov | oF | sr |
cI 62 , 98 |

*£ipunoy Z “ON
eeeeee 4ST ‘sapunoy Zz “ON

*£0(DNCT “UIT PLO

eee eeer eee eee cee

ress puz ‘AApunoy Z “ON
"pug ‘ipunoy g “ON
eee “aS 105 uowul07
peers agT ‘sapunoy ¢ ‘ON
“Sid Zuyjour g ‘on

sere ** 4sBiq PIOD

** 4881q P10 [ “ON
seers Sid Lipunoy Z% “ON
sees > “Q9RIG PIND & “ON
Sure 4s] ‘ew eMey
seeeeee aB103 ¢ puv f ‘ON
seeeee e810} O puv [ ‘on

eeetereeeee “

- soeeeeooes BIOTATBIST
seteeeneeeeeeersees UOABDOBIE
POOP e EMOTE eee eee eeeee B10} s[eis
aiTysf{qieg ‘ureyiey) 389M
eeeereee jood 44qu0g
“+++ doaBuanIg

Aatpng ‘107.1943 'N |
**** ** TOAROORIGT |

* Soipng ivou ‘sy10 M4 sfv'T
*saypng ‘u07.19u10N
senses eee” POABUIE |

| yt
| er

rat
Il

=
=}
L|S
“@S19ASUBL], | Q=nSer-R

‘oqydeiy

-gyisuay, | Caen ete Nene

“UTS,

“soryTenb Burys10 4 “word ioseq *peyddns moym sq

“£417enb jo 19pi9 *sisA[vuy [worulagD

‘woonrg fo sapicy ua (‘qua0 sad F.C 07 %.0) wor-brg osoydsoyg fyzybyug—''] 11AaVI,

"a31y ‘uy pooy

4j08 puv poor

seeveceeeces uoit-31d @ "ON
* Sd Aspunoy | ‘ON
seeeee a1 fapunoy €°ON
Zid A1punoj F ‘ON
seeeeeeeereree aid Aapunoy
eee eee teen eters 3103 foi
** payjauey

*** wost-31d 1°ON “ “
** uoil-3id F ‘ON * "OD UOJ] eTeprea mM

"1%-3 = "DO [hol pooy

4jos puv poory

4
°
im
~
sal
MQ
<
oO
is
o)
mM
)
-
BH
[a=]
i)
a
oO
io=]
i
&

“"UeABqorTT MA ‘oryyeULeH
“ “ec

* 1490/5 ‘soovusng puy yueg

oPrerrer etter) “

poop
MARL

oocoooooooo

“g31q ‘UN

Ivoerto-—
-—Oo=-oo7—

‘ouydesy

“Burgsnig
ogloeds

*syvUTy “sarqtpenb Burysom fruroyy *uorjdL10seq *poyddns woya 4g

“@SLGASUBLIT,

“Auypenb jo s9pi9 “stsf4[eue [eolrmeyD

“wong fo 2psQ Ww (-zue0 sad %.0 topun) u012-b1q o240ydsoyd-uoxT—* YI AAV,
. 4

138 TURNER: THE INFLUENCE OF SILICON

of the series. On examining the average order of quality, we find
that there is a marked resemblance to my own results, but three
specimens are a little exceptional, being in two cases inferior to what
might have been anticipated from their silicon content (Nos. 2 and 5),
while in the other case (No. 9) the quality is improved. The first
named two specimens are highest in graphite, and this accounts for
their weaker character, as an addition of but a few tenths per cent. of
graphite beyond 3 per cent. appears to weaken the iron very much,
In No. 9 a low proportion of graphite and over 1 per cent. of
manganese appears to have partly neutralised the bad effects of 4°25
per cent. of silicon. On allowing for these differences of chemical
composition, we find that the series shows an improvement on adding
silicon up to 1°88 per cent., and then gradually deteriorates with a
greater proportion. This effect is exactly similar to that seen in my
own experiments, except that in the present instance it is somewhat
obscured by variations of other elements present.

II. Slightly Phosphoric Pig-iron.

Passing now to material containing a little more phosphorus, we
find that there are 14 specimens mentioned in the report which con-
tain from 02 to 0°4 per cent. of phosphorus. These are arranged, just
as before, in order of silicon (see Table L, p. 137), and they exhibit
less variation due to silicon than any other class of iron mentioned in
the report. The reasons for this are very apparent on careful exami-
nation.

The two last specimens in the list have been separated from the
others by a line, becanse in these two cases it is evident that too
large a proportion of silicon is present, and the metal is weakened in
consequence, as shown by the low average order of quality. The
proportion, however, is not sufficient to interfere with the working
qualities. But on examination of the other 12 specimens it will be
found that the maximum difference in silicon is only 0°92 per cent.,
which gives an average difference between two adjacent specimens
of only about 0°08 per cent. of silicon. But it will be noticed that in
the same specimens the graphite varies by 1 per cent., and manganese
by 0°87 per cent., while the sulphur and phosphorus also introduce
some irregularity. When we further remember liability of experi-
mental error, the fact that the analyses were made before remelting,
and the very rigid character of the test we are applying, it is not to be
wondered at if the results are irregular. In this series of 12 speci-
mens, we find that the effect produced by variations of silicon is counter-
balanced by the considerable and irregular alterations of the propor-
tion of other elements present. The highest and lowest specimens are

cn a n,n a a a ae

ON THE PROPERTIES OF CAST IRON. 139

almost equally removed from 1-4 per cent., and are of exactly the same
average order of quality ; while the mean order of quality of the first
six specimens is 17, and of the last six specimens 17°6, or almost
exactly the same. From these results we may conclude that within
the range of + 0°5 per cent. from 1°4 per cent. of silicon, irons are of
good average, or very superior quality, but that their relative quality
is liable to be much influenced by other comparatively small differences
in composition.

III. Moderately Phosphoric Pig-iron.

In this class are included all irons containing from 0°4 to 1 per
cent. of phosphorus; and for sake of convenience they are divided
into two series.

a. Containing 0°4 to 0°6 per cent.—Of these there are nine given in the
report, and they are arranged in order of silicon in Table M (p. 140).
In this series we have only a difference of 0°91 per cent. of silicon, while
manganese varies by 0°96 per cent., and hence introduces a disturbing
effect. But as each specimen which contains much manganese is
deteriorated in quality thereby, its effect may be plainly observed.
On making suitable allowance for these cases, we find the effect is
quite similar to that noticed in Table K (p. 137), and also in my own
experiments.

B. Containing 0°6 to 1 per cent.—Of these there are eight given in
the report, and we have them arranged in order of silicon in Table N
(p. 140). It will be seen that Nos. 2, 3, 4, 7, and 8 form part of a regu-
lar series, resembling those noticed in Tables K and M. But there are
three exceptions to the regular order. Of these No. 5 is the highest
manganese of the series, and is consequently of inferior quality; and
No. 6 is slightly inferior, though in this case the reason is not so
plain. No. 1 is also out of order, and at first sight it appears to be
80 to a serious degree, as this is the best specimen examined in
the whole of the experiments recorded in the report. But when we
notice that an increase of less than 0°2 per cent. of silicon would
cause it to take what we may assume to be its proper place in the
series, it will be seen how very close the general agreement must be.

IV. Phosphoric Pig-iron.

In this class we include all specimens containing upwards of 1 per
cent. of phosphorus; of these there are 10 given in the report. They
are arranged in order of silicon in Table O (p. 141), and strikingly
exhibit the effect we have previously noted in the other series. The
specimens highest and lowest in silicon are both inferior, while the best

L 2

Zz
©
i)
=
a
ica
o
<3)
12)
a
>
Z
Q
ee]
=

TURNER

“ysry ‘uy

e[qe19[0],
“

2191910],
poor)

*syAVULOY

“Burysnig | -

“sorqyenb Surylo Mm

IWwOomwNny
a

310 O

coooooooco

ut
Soo

Gl.

£¢-1
It-T
OF-T
re. 1
92-1

cooooooeo

Ot
o=-ooco

* @R105 £013 ¢ ‘on

*eB105 Bu0.148 F ‘ON
seseeeeseee F104 LOIN)

seteneeeeseeeeeees aemtG JOH

OYB on[q g “ON

sree Sid Suyjow £

eatysiqiod ‘wielrey 19M

seeeereseoeeeerragmtG DIOD |" ITH 499g ‘sy10M [9407]
“ oe

eeeee “

“-"er1yskqiog ‘UeTey IS8M
" TITH Aopionig ‘sys [2A0'T
serves orrushqiag ‘Aaiming
“"orrgshqiog ‘WRITeY S90

"MOISIO],

*@SIDASUBLY, |

‘aqiydein

“Auyenb jo 19pio

*sisf{[eue pwormayyD

"mond oseq

*poyddns moya fq

‘monty fo sapic

ur (‘quan sad

01.9.0) wou-brg nsoydsoyg

hjaywsapoyy—' NX 14V J,

"ysIy “UW
‘y3ty “UW

“4UsIy
onqdess pus ‘ug

0-81

¥- LE

8- OF

“ g. &2

pooy | Z- ce
O1GBs9[oOL | Z- LI
“| 49

poor
pood £194

0- 2
0-21

“syivUley

*soysyenb Zuryiom Puvsyy

It-0
gc-0
g¢-0
2F-0
£r 0
1t-0
8t-0

tF-0
8g-0

Lz-0
aoRdy
6-0
28-0
96-0
¢-0
98-0

18-1
GL-T
£9. 1
8g. 1
8F- 1
98-1
TZ 1

+6-0
g¢-0

9I-T
96-0

“-puz, ‘sapunoy | “ON
coces +*@B40§ F “ON |

sereees OF105 AOIN)
481 ‘Aapunoy | “ON

seeeeereeees K IDUNOY 1°ON

“98105 Buoig
seeeeenenenee QB OT

seeeeeeseO D107 hoid ¢°ON
eee eeeeeeeeee aid sapunog

. és |

seeeeeers srereees wg AIBISK
sereeees Korpny ‘UH PIO
“seers AOIPNG ‘pwoH A1Vd
eoece rem atA[BIsk
* £oypnq “uojayION |
** a1pN ‘tH PIO
seeeeeeeeeees Saipng ‘peony wed |

cee ‘ao119q19N
eee ‘prey ye

*eTIsUay,

*@S19ASUBLY, |

‘g | ‘oudurp

“£y7enb jo 19pig

“sisfwue peormegD

*uondoseg

*perddns woya sg

‘uoonig fo sapic ur (quad sad 9.9 03 F.0) vour-hig nsoydsoyg hyaznsepoyy— 414],

z
2
-
mM
<
s)
fe,
iS)
7]
=|
=
ca
fal
Bu
S
x
A
<3)
ce)
a)
ZA
)

*s911089 JO "gq pur
onydais 8 4saMoy

“MOT
“Ad ‘G81q eyydeip

qmowia
pooy

91qQB19/0],

e[qesa[oy

poor
s[qeia[oL

*sqIVUIOY

*sorytpenb ZuyyioM

*9S10ASUBLL |

e9-%
€L-%

11-2
9f-%
26-1

90-T
| $2.0

soorrees Furqiour ysuy ‘Biq
“+ £Ipuuojs Z “ON

oteeee u0ll-34q
eeeeee - Sapunoy Z ‘ON
sevevececcooecoes CIR I ‘ON
pocccccococe Lameengg € ‘ON
becee eorongnata age 5 “OE
“qsu[q 104 ¢ "ON

***£ipunoy | ‘ON
*** paveysnoj Z ‘ON

+e TODAGM Iva ‘piopAoH]
**** OIQSO[PPIT ‘HUB YMoy

*‘pioyeig ‘N ‘opepueplop
‘asiysAqieqg ‘fsa[4019ng
O1QBUI[eM ‘pug 3seq

“aaqysdqueg ‘sojseqng

“ “ ‘

sees

"SHO M UOIT 090}

*"-artysiquog ‘sas0qINg
* O1Q82/PPIA “Wuvg Ynog

‘g | ‘aiudeiy

*Syryanb jo sopsg

*sisf|BUB [BOT MIDYO

‘uoTydp10seq

*porddns woqa fq

‘mooyrig fo sapagQy ur (“7Ua0

“ad { 4aa0) uom-big nsoydsoyg—'Q ATAV,

142 TURNER: THE INFLUENCE OF SILICON

specimen contains 1°57 per cent., and agrees remarkably with the
recommendation of Mr. Wood for Cleveland iron, which is of a
similar character to the iron in this series. In a note incorporated
in my previous paper (Trans., 1885, 913), Mr. Wood recommended
1°5 to 1:75 per cent. of silicon for production of strong foundry iron
from Middlesbrough pig. In Table O(p. 141), there are two specimens
which are slightly irregular (Nos. 2 and 8), each of them being of
rather superior average quality. This is explained by the fact that
they contain the least phosphorus in the series; and also that in
No. 2, where silicon is low, graphite is high ; whilst in No. 8, where
silicon is high, we have the lowest graphite of the series. These facts
are quite sufficient to explain the slight irregularity noticed, and the
series is then of remarkable uniformity.

We have now come to the end of the report itself, which dealt with
53 specimens; 50 of these we have considered ; 2 were too hard
to turn, and have been previously mentioned ; whilst in the remaining
case no analysis was given. In the appendix, 17 specimens are
mentioned, 7 of which are foreign irons, and are given in Table F,
p. 133; 2 were prepared by patent processes, and these are not sufli-
ciently numerous for comparison; whilst of the remaining 8 we have
only 6 analyses. The results of these are given in Table P, the
specimens being arranged, as usual, in order of silicon.

In these experiments, the manganese was not determined, and the
mechanical tests were also incomplete, but the same kind of effect is
noticed as in my own experiments, the best results being obtained
with a moderate amount of silicon. The very sudden decrease in

Taste P.—English Irons.

Analysis.

Locality, &c. . / Remarks.
. | Gra- :
| phite.| 5: | P-

. Ulverstone ... "59; 2°83 |0°03/0-1¢ Too hard to
turn.

. Bowling...... 0° 99 [0-050 -5¢
3. Bowling, No. 2,
2nd sample.. | : 0 *03)/0 *4¢

. Bowling, No. 3, |
2nd sample.. |

. Bowling, No. 2, |

0 -04/0 *52

lst sample .. 0-060 56
- Bowling, No. 3,

Ist sample ..

0 ope

> ae he fe DS ©

Oo 2 Ge -

i EE ee ee

UN THE PROPERTIES OF CAST IRON. 143

tensile strength noticed in No. 5 is probably connected with the large
proportion of graphite present.

We have now considered every specimen mentioned in this most
valuable report, and my endeavour has been to bring forward the
evidence in such a manner that all who are interested may be able to
judge of the validity of my conclusions. My friend Mr. John Spiller
allows me to mention that he has carefully gone over my tables, and
considers that in their preparation the report has been honestly dealt
with, and, in the main, he coincides with my conclusions.

Sir W. Fairbairn’s Experiments——In the Report of the British
Association for 1853, an account is given of some experiments by Sir
(then Mr.) W. Fairbairn, undertaken to ascertain the effect of
repeated meltings on the properties of cast iron. As this question is
one of very considerable practical importance, I shall venture to deal
with these experiments in detail. The results have been often quoted
by different writers, and the mechanical tests appear to have been very
careful and complete, although it is to be regretted that the chemical
analyses were not so fully performed. In Table Q (p. 144) is given a
summary of the most important results obtained; and in addition to
what is included in the report, 1 have added the calculated tensile
strength from Box’s “Strength of Materials,” and also the probable
amounts of silicon, sulphur, and carbon contained in the best speci-
men, as calculated from the other analyses given, and the average
rate of change of composition.

The iron used was Eglinton No. 3, hot blast, which was remelted
18 times. We cannot here enter into details as to the shape of test
pieces, and method of operation ; but it will be seen from Table Q the
metal was at first slightly deteriorated by remelting, that afterwards
it was much improved, reaching a maximum at the twelfth fusion,
but on further melting it again deteriorated, and became very inferior
indeed after the fifteenth fusion.

These effects have generally been attributed to remelting, as if there
were some particular charm in the process of fusion capable of improv-
ing the quality of the product. But Dr. Percy, in speaking of these
experiments says, “As fusion was effected in a cupola (comp. note,
p. 147) with coke as the fuel, it is obvious that in the course of re-
melting the iron might have suffered important chemical changes in
composition, which may greatly modify its properties. It is hardly
conceivable that simple melting, without the addition of anything to,
or the subtraction of anything from the metal, should produce any
decided effect.” (“Iron and Steel,” p. 862.)

Dr. Percy’s suggestion of “important chemical changes,” during
remelting, is fully borne out by an examination of Table Q (next page).
It will be seen that silicon increases from 0°77 to 2°22 per cent., the

Zz
)
2)
_
=
—_
mM
fx
o)
Le
'S)
PA
&
fs
4
_
aa
=
B

TURNER

“UOTFISOdUIOD OTGVqorg 4

‘ord 3. ‘spuroyeypy jo 4

usI14g ,,

‘xog “iL

Moay USHVy St W43u0148 STISUS} POFB[MOTVO OTL],

‘2ORp.ANS OTOYM O44 10A0 AUOATIS PUB OFIY AA

“prey

fiona {oaqueo ut ,,¥ ydeoxe ‘oy As0aTig
*19uLOD YoRe

qe ageym fusaps Soxyaco ut Load 4ysvy

96T-

9TT-
998.

0-88

| 8- SFT

9LF-0

L-@1&

ess. Z | GL-.€

Cl.O | 22-3

*S1BG OFUT UNA OF F148 009 Sulog Uoal oy} ‘Oanjrez B SUM SurqoU YIZT OTT,

¢-0L
L-9L

S- 86T
9- 8&6

999-0
9-0

€- 168

T- TLE

oss: Z

8tz.L

“eT aq
pus pavy yeqymouros { ford pouread oso[p
*spuqysdu0 aepn3ue
divys yyim ‘uomt poured osojo Su0yg
*‘£oad yysty foangxo4 peureas Ljoug Bucayg
*QULBS OY} OS1M.19990 ‘ANO[Oo AO4Vq TI] LOY VY
“UOJ
-tan Ajwossod ‘pavy s0yyer ‘uread wat
“euLes Inoj0d “youduIOD pus WAOFIUN OL0TT
*OLOJO SB WILOJ pUB INOTOD
*‘posvauout ssoupaey pus fy1su0q
‘oues 044 couvivedde ‘4308 [eyoy_
"103.18
speyshio solmojyur {ysmyq {fosysny os0zT
“eL0JOq 8B INO]OO ‘Aapisy [ey

PST-6

Pa.
Tg:
TZ.

OF- OT

L0- O01
Lt8-6
c¢9-.8
6FE-8

It
er
Il

IST-8
469-4

6- 96
(oT 004)
0-99
T. &Z

o16

FPOT
SIT
990T

126
css
Ty8
99
64S

89
063

| asee re

¥-€09

8-FE9
I- 269
6. 19

6-999
S- 9S
€- 16V
I. 6hP
L-88P

9- 189
v-ely

08s. Z

PEI-L
O9T- £
S12

SOT. 2
ZOI- 4
C20. L
648-9
124-9

3r8-9
8&6-9

4+(88-0) (40.2)

92-0

09-0

"OULBS OY} OSTM19Y4IO ‘arYSN] S87]
*qoodsoa L190 ut ures oy} ApABaN
"914X094 UI UsozIuN L194 ‘pouresd s0],)

Tse. 4
L1Z-8
LOS- 6

-96¢
- 089
SOL

warn |2o SCONMM BOF

9- 10P
6-1hP
0: O69

988-9
046-9
696-9

rN oO Fw OMDAD

*syIVUloy

‘ut ‘bs
aod suo,
"qy3u0838

o[tsu94

#P27B8[NITB_

‘uorssord
-M109 03
e0uBysIseyy

*goudunt
4sIsoa
04
1OMOg

“sOqOUT UT
uoTpepep
eyeuriyyn
uBsoyy

‘ul ‘bs
aod ‘sq
ar 4qZ10m
Suryvorq
uveyq

*“£41avid
oywedg

jo oSvyus010g7

(48 “d ‘gceT
“sodey ‘oossy ystyi1g) wour ysnQ fo sansodorg ponunysezy ey} uo shuryepy pazveda.s fo yoofig ey,—? AAV],

+ Probable composition.

=
's
=
=
3
-—
=
7)
~~
3
=
a
—
°
«
=
~
wo
=
=]
+
=
~
2
.

* The calculated tensile strength is taken from T. Box,

ON THE PROPERTIES OF CAST IRON. 145

increase appearing to be of a fairly uniform character. The amount
of sulphur varies irregularly from 0°26 to 0°75 per cent.; while the
proportion of carbon varies, not very regularly, from 2°3 to 3°75 per
cent.

Under these circumstances it is evident that the chemical changes
are amply sufficient to account for the effect produced, quite apart
from any change due to remelting simply. In order to make this
clear, the table is divided into three parts by horizontal lines, and
these parts we will consider in order.

1. The three first meltings produced a deterioration in the
properties of the metal. The cause of this cannot be ascertained with
certainty, as we have no analysis at the end of the third melting.
But it is quite possible that the increase of sulphur, of which we find
evidence in the next analysis, took place about here, the sulphur
being derived from the coke employed.

2. In the 11 following meltings, we find the sulphur and carbon
somewhat irregular, but silicon appears to have been added slowly
and with considerable regularity, and its effect is seen to be exactly
the same as in the experiments to which previous reference has been
made, including the Woolwich experiments, Mr. Wood’s recent work,
and those 1 have given in Parts I and II of the present paper.
In this case, the influence on the mechanical properties of the
metal is very regular and well marked, the maximum improve-
ment being produced by about 2 per cent. of silicon. This amount
is a little greater than that found in the other experiments,
but the iron operated on contained more sulphur than in any
previous case; and as sulphur and silicon, in due proportion, appear
to have almost opposite effects, it would be naturally expected that
the most suitable proportion of silicon would be higher than usual.

3. In the four last meltings, the quality is very inferior: the most
suitable proportion of silicon has now been exceeded, as we saw
previously ; but it would be quite unfair to attribute the whole of
the sudden decline in quality to this cause. It will be seen that the
last melting contained no less than 0°75 per cent. of sulphur, while the
tenth melting had only 0°26 per cent. A careful consideration of the
properties of the metal will show that sulphur considerably increased
about the fourteenth and fifteenth meltings, rendering the metal hard,
white, and brittle; and these changes were accompanied by a very
sudden decrease in strength.

It will be seen from this brief survey, that the effect attributed to
remelting may be sufficiently explained by a consideration of the
chemical changes which took place. It may be added that these
results are quite inexplicable on the supposition that “silicon is
always bad ;” but on the other hand, when due allowance is made for

146 TURNER: THE INFLUENCE OF SILICON

the disturbing influence exerted by sulphur, they afford strong
confirmation of my own results.

It is not contended that remelting, when unaccompanied by
chemical change, is never beneficial. It is well known that castings
made directly from the blast furnace are unsatisfactory, and in this
case probably the iron would be improved by remelting, even if it
could be accomplished without chemical change, as it would ensure
greater uniformity in the metal. In this respect the common practice
of casting from a cupola, especially with suitable mixtures of iron, is
to be commended. But any attempt to further improve the quality
by remelting is not wise. It has been shown that some metal is
improved up to the twelfth melting ; others to the eighth; some only
to the fourth, whilst some cast iron is deteriorated by even one
melting. This will obviously depend on the character of the metal
and the changes which take place during remelting, such as
alterations in silicon, sulphur, carbon, &c. It is evident that these
changes may be produced more readily, and with greater certainty,
by judicious mixing, than by the laborious and expensive process of
remelting.

It was pointed out by Robert Mallet, so far back as 1856 (‘‘ Construc-
tion of Artillery,” pp. 13 and 218), that the results obtained in
Fairbairn’s experiments did not depend upon remelting alone; but
at that time our knowledge of the influence of the various constituents
present in cast iron was insufficient to enable either Mallet or Percy
to point out the precise reason for the changes, though they correctly
indicated the directions whence the alterations proceeded.

In addition to the experiments to which I have previously referred,
there are a number of other facts which strongly support the conclu-
sions drawn from my own experiments as to the influence of silicon in
cast iron. Among the more important of these may be mentioned
Stirling’s method for producing toughened cast iron, by mixing soft
grey pig with wrought iron scrap in suitable proportion; Price
and Nicholson’s patent cast iron, produced by mixing white and grey
iron in suitable quantities; and lastly, the more recent use, chiefly in
America, of steel scrap as a means of strengthening cast iron. The
advantage of these methods is undoubted in many cases, but it is not
to be wondered at that unsatisfactory results were often obtained
when silicon was regarded as being a noxious impurity under all
circumstances. As a consequence, the proportion of silicon was often
lowered in metal which already contained too little, with the result of
rendering the iron both hard and weak. This part of the subject has
a more practical bearing than the purely experimental part, and s0
will probably be considered more in detail elsewhere. But it may be
added that in each of the cases mentioned the object aimed at was not

ON THE PROPERTIES OF CAST IRON. 147

to produce a steel, but a cast iron of special quality, the carbon and
silicon present being generally reduced by only one-eighth to one-
fourth of the original amount.

On examining the various tables previously given, and selecting
from each the specimen which apparently contained the most suitable
proportion of silicon, we have the following result—

Silicon,
per cent.

. English irons in the Appendix (Table P).... 1°03
Containing 0°4 to 0°6 per cent. P. (Table M).. 1°21
. Foreign irons in the Appendix (Table G).... 1°33
Average of six best specimens (Table H).... 1°39
Containing 0°6 to 1 per cent. P. (Table N).... 1°41
. Containing over 1 per cent. P. (Table O).... 1°57
. Containing under 0°2 per cent. P. (Table K).. 1°88
. Sir William Fairbairn’s experiments (Table Q) 2°04?

Average 1°48

It will be seen that this average amount varies but little from that
which I have previously suggested (1°4 per cent.) as the most suitable
proportion for strong iron (Trans., 47, 912). And, when discussing

the results in detail, we have already seen the reasons which tended
to make the amount somewhat irregular, where any considerable
variation from the average is observed. For my own part, on
considering the facts now brought forward, together with the results
of Mr. Wood’s work and my own experiments, I regard it as
abundantly proved, that silicon, when present in suitable proportion,
is a true friend to the iron-founder, since, with no extra expenditure,
a stronger, softer, sounder iron may be produced.

In bringing my paper to a close, I would express my belief that by
a knowledge of the constituents of the cast iron employed, and a
suitable mixture of material to produce the best result, the average
strength of cast iron in this country could be very considerably
increased. At the same time good working qualities could be
regularly obtained, while no extra cost in production need be
incurred,

Note by the Author.—In the discussion which followed the reading
of this paper (Abstracts of Proceedings, p. 135), it was mentioned by
Professor Unwin, who took part in Sir William Fairbairn’s experi-
ments, that the remelting was performed in an air furnace, and not
in a cupola, as stated in the extract from Dr. Percy given in the text.
I am pleased to be able to make this correction, as it explains why so
many meltings were possible before the metal suffered seriously in

148 TURNER: THE INFLUENCE OF SILICON.

properties. Not being able to find any record of the methods adopted
for remelting, I naturally accepted the statement of so excellent an
authority as Dr. Percy. Otherwise, however, my conclusions are
unaffected by this slight error. The fact remains that important
chemical changes occurred, and that these changes, which were
probably gradual, are amply sufficient to account for the alterations
in the character of the product, quite apart from any virtue in
remelting per se. In the discussion it was stated that “the
change in properties on remelting has always been attributed to the
change from graphitic to combined carbon;” and remelting is
commonly believed to produce this result. It may, however, be
pointed out that remelting does not always increase the amount of
combined carbon; it has occasionally an entirely opposite effect.
Thus, under certain circumstances, white iron may be converted into
grey by remelting, quite apart from any effect due to slow cooling.
The probable causes for the belief that remelting hardens iron are two-
fold. First, the fact that the fuel (coke) oftens contains a consider-
able proportion of sulphur; and secondly, as foundry iron is a soft
grey, seldom containing more than 0°5 per cent. of combined carbon,
we have practically only one kind of change possible, namely, a
reduction of graphitic carbon, for obviously it cannot under these
circumstances be increased.

It is not sufficient to say that the graphitic carbon is altered. We
want to know wiy this is altered. It has yet to be proved that
remelting, apart from chemical change, is capable of producing this
effect. On the other hand, we do know that addition of silicon to a
white iron produces yrey iron, as shown in Part I; while the addition
of sulphur to a grey iron produces a mottled or white product.
Hence, when no other constituent is present in unusual quantity, we
can predict with tolerable accuracy the relative proportion of
graphitic to combined carbon in a sample of iron, when the propor-
tions of silicon and sulphur are known. Nowsince for this prediction
we do not require to know the number of times the iron has been
remelted, but only the chemical analysis of the product, it is obvious
that the alterations in the proportion of graphitic carbon are due to
chemical changes and not to the mere operation of fusion.

It is for this reason that I have not referred to the proportion of
carbon in Sir W. Fairbairn’s experiments. It will be found that in the
original paper mention is made of the unsatisfactory methods by which
carbon was then estimated, and no distinction is made between
graphitic and combined. But in the variations of sulphur and
silicon we have an index to the variations in graphitic carbon; and in
these experiments we have, as it were, a struggle for supremacy
between sulphur and silicon. In the second portion of Table Q,

a. af A at fh 2a OU eeeelUCirK Rl elCOF

SEPARATION AND ESTIMATION OF ZIRCONIUM. 149

according to the analyses, we have sulphur low, while silicon gradu-
ally increases, and with this increase we find a result exactly like that
noticed in my own experiments. These results therefore confirm my
previous conclusion that a suitable proportion of silicon exerts a
beneficial influence on the properties of cast iron.

XVI.—On a Method of Separation and Estimation of Zirconium.

By G. H. Barter, D.Sc., Ph.D., Assistant Lecturer on Chemistry at
the Owens College.

The Preparation of a New Oxide, Zr,0s.

Last winter, whilst making an analysis of the mineral koppite,
in Professor Bunsen’s laboratory in Heidelberg University, it was
found that one of the precipitates, containing ferric oxide, yielded a
notable quantity of zirconia on treatment with tartaric acid and
ammonium sulphide. Thinking it might consist, at least in part, of
titanic acid, this precipitate was redissolved and hydrogen peroxide

was added to the acid solution; no trace of the reddish-brown
coloration characteristic of titanic acid was observed, but a white
bulky precipitate was formed. This was well washed, and heated
gently with the addition of hydrochloric acid and potassium iodide;
iodine was liberated, indicating that the substance was a higher
oxidation product of zirconium. Toconfirm this, pure double fluoride
of potassium and zirconium was taken, and having been converted
into the sulphate was treated in the same manner with hydrogen
peroxide ; the precipitate obtained agreed in all respects with that
already observed in the examination of the koppite.

About the same time, though I did not see it till months afterwards,
a paper by Cléve appeared on the action of hydrogen peroxide on the
rare earths (Bull. Soc. Chim., 43, 53), showing that he had obtained
a peroxide of zirconium having the composition ZrOQ;. This he pre-
pared by precipitating with ammonia a solution of the sulphate con-
taining hydrogen peroxide, and washing the precipitate by decantation.
Such a method is open to the objections that being precipitated par-
tially by the hydrogen peroxide and partially by the ammonia, the
product consists of a mixture of hydrate of zirconia and the higher
oxide ; that with such a substance as zirconia it is doubtful whether
the whole of the excess of hydrogen peroxide can be washed out by
decantation ; and, lastly, that under the circumstances, as pointed out

150 BAILEY: METHOD OF SEPARATION AND

by Cléve himself, notable quantities of nitric acid are produced, no
doubt arising from the reaction between the ammonia and the
hydrogen peroxide. On reading his paper, therefore, in view of cer-
tain results which had been arrived at already, it appeared to me
desirable to continue the work, and ascertain whether it was possible
by the addition of hydrogen peroxide alone, to prepare an oxide of
definite composition, and whether a complete separation of zirconium
might not be effected by its means. A quantity of potassium zir-
conium fluoride was prepared from zircons by Wohler’s method. The
melt was powdered, digested with water and filtered hot, the process
being repeated as long as any of the salt entered into solution; the
double fluoride crystallised out on cooling. The insoluble residue
contained iron and strontium in tolerable quantity, whilst the zir-
conium salt was somewhat yellow from the presence of a small
quantity of iron. At this stage, a separation of the iron was not con-
sidered necessary. The salt was treated with excess of concentrated
sulphuric acid to expel the hydrofluoric acid, and the solution after
dilution with water was precipitated directly with hydrogen peroxide.
The precipitate thus obtained was quite free from iron, and after
washing out any free hydrogen peroxide and potassium salts it was
preserved in the moist state. Three months afterwards this compound
was subjected to a quantitative examination.

A portion of the moist oxide was introduced into a small flask and
distilled with hydrochloric acid, the product of the reaction being
passed into a solution of potassium iodide in the manner adopted in
iodine titrations. The amount of iodine liberated was estimated by
Bunsen’s method, using an approximately decinormal solution of
iodine and weak sulphurous acid. At the conclusion of the experi-
_ment, after the contents of the flask had been well boiled, and no
more iodine was set free, the zirconium was precipitated as hydrate
by the addition of ammonia, independent determinations having
shown that the zirconium undergoes complete precipitation in this
way, and that none of it distills over into the solution of potassium
iodide. The precipitate, washed and ignited, gave the amount of
zirconia. In all cases this was snow-white, and showed no trace of
iron. The following is one of several concordant observations made

in this way :—
Gram.

Iodine set free, equivalent to 17°2 c.c. “ — 02494,
standard solution .
ZrO. found 0°2520
Zr calculated from ZrO, ............ = 0°18602
O 0:06598 .
9 =0°08181
O iodine liberated .. = ooises | "7

ESTIMATION OF ZIRCONIUM. 151

From these numbers it appears that the peroxide is Zr,0;. The
percentage composition is-—
Found. Calculated.
69°455 69°284
30°545 30°716

O = 15°96, Zr = 90.

Cléve’s results lead one to anticipate an oxide of a somewhat
unstable and ill-defined character, and so it might have come about
that after three months’ keeping, the compound, though answering so
well to the composition Zr,0;, had yet lost oxygen during this time.
A fresh portion was therefore precipitated with hydrogen peroxide,
and after being well washed and the water removed at the filter-
pump, it was subjected to analysis in the same way as the previous
specimen, the following results being obtained :—

Gram.
Todine set free, equivalent to 6°9 c.c. “ = 010003
standard solution
ZrO, found = 0°1032
Zr calculated from ZrO, = 0°07618
O - = 0°02702] _.
O = 000631 } aaa

”

The percentage composition is—

a result agreeing with the previous one, and indicating that the
oxide is of a very definite and stable nature. The slight oxygen
deficiency that does occur may be due to a retention occasioned by
the unavoidable dilution of the hydrochloric acid used through the
oxide being in the moist state. This is rendered the more probable,
since with a more dilute acid the oxygen was 0°48 per cent. low, and
using hydrochloric acid diluted with its own bulk of water, the
deficiency was 0°85 per cent. The composition of the oxide obtained
by precipitating a solution of zirconia in dilute sulphuric acid by
means of hydrogen peroxide, is therefore Zr,0;. It is worthy of
remark, that whereas zirconia when freshly precipitated dissolves in
the minimum quantity of very dilute sulphuric acid, this higher
oxide is much less readily acted on, and indeed by treatment with
cold dilute acid, a separation of the two oxides may be made. In the
formation of such an oxide, therefore, as in many of its salts,
zirconium shows distinct proclivities towards the nitrogen group of
elements, and more particularly towards niobium, tantalum, and
vanadium.

SEPARATION AND ESTIMATION OF ZIRCONIUM.

Zr,0; as a means of Separating and Estimating Zirconium.

The preparation of zirconia free from traces of other elements is a
matter of considerable detail, and where iron, titanium, niobium, and
some other of the rarer elements are present, it is only effected with
difficulty. For the estimation of zirconium in minerals which
contain but small quantities of this element, there is no process
that can be considered satisfactory. At the same time, it is by no
means improbable that zirconium is much more widely distributed in
the mineral world than we have as yet direct evidence of. Hydrogen
peroxide presents us with a means of separation. A very dilute solu-
tion of this reagent does not effect the precipitation of zirconium,
but, with a moderately concentrated solution, the precipitation is
complete. Cléve has pointed out that iron and alumina are not pre-
cipitated by it. Titanium gives a coloration, and forms a higher
oxide, but this remains in solution, and the pentoxide of zirconium
precipitated from a solution containing titanium salts was found to be
quite free from titanium. Niobium and tantalum also appear to give
no precipitate with hydrogen peroxide, nor do tin, silica, or indeed
any of those elements which usually occur along with zirconium, and
are difficult to separate from it.

The examination of the reaction is not yet complete, the resulis
however, obtained up to this time are summed up briefly in the
following statements. Hydrogen peroxide of such a strength as to
yield 120 vols. of oxygen on heating (Thénard’s strongest peroxide
gave 475 vols.) precipitated the Zr,O; immediately and completely
from a solution of zirconia in sulphuric acid, containing consider-
able excess of the acid. If its strength were such as to yield only
20 vols. of oxygen, the precipitation occurred after a few seconds, and
was not quite complete; on standing some time, however, in a closed
tube, it was found that even with this weaker solution every trace of
zirconium had been precipitated.

By the use of hydrogen peroxide it is therefore possible—

(1.) To prepare most readily pure compounds of zirconium.

(2.) To separate even small quantities of this element in minerals.

(3.) To make a quantitative estimation of zirconium in a direct
and rapid manner. I propose to continue the examination of the
reaction, and hope somewhat later to give such further details as
may be of interest.

XVII.—Notes on an Analysis of Koppite.

By G. H. Bartey, D.Sc., Ph.D., Assistant Lecturer on Chemistry at the
Owens College.

APPENDED are the results of the analysis of a specimen of this
mineral from the Kaiserstuhl referred to in the previous paper. The

analysis made by Knop and Rammelsberg is given in the second

Knop and Rammelsberg.
62°18
3°00
6°69
180
16°00
0°40
5°58
4°23
(6°23)

(1.) The whole of the work of solution and digestion was done in
platinum vessels. Forty times its weight of concentrated sulphuric
acid was used for dissolving the mineral, a clear solution being
obtained when the latter was added in small successive portions to the
boiling acid. For such analyses, sulphuric acid is to be recommended
in preference to hydrogen potassium sulphate.

(2.) The titanic acid was estimated by Weller’s colorimetric method
(Ber., 15, 2592) with hydrogen peroxide.

(3.) Rammelsberg, in another analysis, gives the percentage of
fluorine as 2°9, and this is probably more typical; at any rate the
specimen examined by me contained very little fluorine indeed.

(4.) It is well known that substances such as niobic acid and
zirconia readily pass through the pores of the filter-paper when
filtering under pressure, and to avoid this it is usual to add to the solu-
tion ammonium acetate. The addition of this salt is some trouble, and
increases the risk of introducing impurity in the course of the analysis.
It was found possible to filter without its addition by using the following
device :—A third tube or a T-tube was introduced into the filtering
arrangement, and partially closed by the pressure of a screw clamp on
thick pump tubing attached to the tube. Supposing such a valve

VOL. XLIX. M

154 PERKIN AND CALMAN: BENZOYLACETIC ACID

admits air at 200 mm. pressure, then the pressure at the pump
cannot rise beyond this point. So long as the pressure at the pump
continues to rise (as it does all the time in the ordinary process of
filtering) the filtrate passes through milky; when it reaches the limit
determined by the valve, it begins to go through clearer, and
presently quite clear. The milky filtrate may now be poured back
and the filtration proper commenced. No trouble will be experienced
in getting a clear filtrate. If the valve, however, is closed somewhat
and the pressure at the pump rises, the oxide is again forced through
and the filtrate becomes milky. The same has been found to hold
with barium sulphate. Most of the work embodied in these two
papers was done under the direction of Professor Bunsen, and I am
indebted to him for many valuable suggestions and for his ever
ready advice.

XVIII.—Benzoylacetic Acid and some of its Derivatives. Part IV.

By W. H. Perkin (Jun.), Ph.D., and A. Caiman, Ph.D.

Ix the three previous parts of this paper, the preparation and pro-
perties of ethylic benzoylacetate and a few of its derivatives were
described, and the close resemblance between these and the ethylic
acetoacetate derivatives pointed out. As, however, there were still
several doubtful points which required clearing up, a further exami-
nation of the derivatives of benzoylacetic acid was undertaken. The
results of this investigation we now lay before the Society. Before
proceeding with this research, it was thought probable that if methylic
benzoylacetate were substituted for the ethylic salt as the starting
point, more crystalline products might be obtained; it being a
well-known fact that methylic salts crystallise more easily as a rule
than ethylic. ,

The mode of preparing methylic benzoylacetate is exactly analogous
to that of the ethylic salt (Trans., 1884, 170), the only difference
being that methylic phenylpropiolate* must be substituted for ethylic
phenylpropiolate.

The crude methylic benzoylacetate produced by the action of
sulphuric acid on methylic phenylpropiolate is a brownish oil which
does not solidify at 0°. To purify it, it is shaken with fairly con-

* This is easily produced by treating phenylpropiolic acid with methyl alcohol
and hydrogen chloride. It is a colourless oil.

AND SOME OF ITS DERIVATIVES. 155

centrated caustic potash solution and animal charcoal, and filtered,
the temperature being kept as low as possible by adding ice. The
filtrate, which should be perfectly clear and free from oily drops, is
then mixed with ice and acidulated with dilute sulphuric acid; this
causes the precipitation of an oil, which is separated from the acid
liquid by agitation with pure ether. The ethereal solution, after
being washed with dilute sodic carbonate, is dried over calcic chloride,
filtered, and the ether distilled off; an almost colourless oil is left
behind, which after standing for a week over sulphuric acid in a
vacuum gave the following numbers on analysis :—

0'1800 gram substance gave 0°0913 gram H.O and 0°4440 gram CO,.
Theory.
Found. C,H;-CO-CH,-COOCH;,.
67°27 per cent. 67°41
5°62
26°97

Methylic benzoylacetate appears to be more stable than the ethylic
salt, it is certainly not so easily decomposed by distillation.

On the addition of ferric chloride to the alcoholic solution, it gives
the same beautiful violet coloration as the ethereal salt. If methylic
benzoylacetate is added to an ethereal solution of sodic ethylate, and
the whole allowed to remain for a short time, a thick white precipi-
tate separates which is the sodium-derivative. In order to purify it,
it is collected as rapidly as possible, well washed with dry ether, and
dried over sulphuric acid in a vacuum. The following numbers were
obtained on analysis :—

0'3470 gram substance gave 0°1206 gram Na,SQ,.
Theory.
Found. C;H;-CO-CHNa-COOCH;.
11°14 per cent. 11°50 per cent.

Prepared as above this sodium-derivative is a white amorphous
powder easily soluble in water and hot alcohol.

It is extremely stable, much more so than the sodium-derivative of
ethylic benzoylacetate, and behaves much like a salt of an acid. On
acidifying its aqueous solution with dilute sulphuric acid, methylic
benzoylacetate is reprecipitated together with a certain amount of
benzoylacetic acid.

When treated with benzyl chloride in alcoholic solution, it gave a
product, which after repeated fractioning boiled fairly constantly
between 250° and 255° (50 mm. pressure). It gave the following
numbers on analysis :—

0:214 gram substance gave 0°5955 gram CO, and 0°117 gram H,0.
mM 2

PERKIN AND CALMAN: BENZOYLACETIC ACID

Theory.
Found. O,H;-CO-CH (C;H;)-COOCH,.
75°89 per cent. 75°00 per cent.
625 ,,
18°75 "

This compound, therefore, was probably methylic benzylbenzoyl-
acetate. A few other derivatives of methylic benzoylacetate were
also prepared, but as they did not crystallise, and as the yield was in
every case inferior to that obtained when ethylic benzoylacetate was
used, the remainder of this research was carried out with the ethylic

salt.
Ethylic Methylbenzoylacetate.

To prepare this compound, 11 grams of sodium were dissolved in
120 grams of absolute alcohol, and a mixture of 90 grams of ethylic
benzoylacetate and 80 grams of methylic iodide were slowly added,
taking care to cool well during the operation. The clear solution was
then heated at 100° for three hours in a soda-water bottle, when the
reaction was complete. In order to isolate the product, water was
added, and the liquid extracted twice or thrice with ether. The
ethereal solution was then well washed with water, dried over
potassic carbonate, and the ether distilled off; the dark-brownish
coloured oil thus obtained was further purified by fractioning under
a pressure of 300 mm.

By the first distillation, almost the whole went over between 230—
240°, and after repeated refractioning a colourless oil of aromatic
odour was obtained, boiling constantly at 235° (300 mm). It gave
the following results on analysis :—

0°1227 gram substance gave 0°3135 gram CO, and 0°788 gram H.0.
Theory.
Found. C,H,-CO-CH (CH;,)-COOC,H;.
69°69 per cent. 69°90 per cent.
6°80 "
2340 4

This compound is therefore ethylic methylbenzoylacetate. It boils

at—
235° under a pressure of 300 mm.

226—227° under a pressure of 225 mm.

Its alcoholic solution gives no coloration with ferric chloride. If
sodic ethylate is added to its ethereal solution, a thick white sodium-
derivative is precipitated, which is decomposed on the addition of

much water.

AND SOME OF ITS DERIVATIVES. 157

In order, if possible, to obtain the free methylbenzoylacetic acid,
the pure ethereal salt was mixed with pure concentrated sulphuric
acid, and allowed to stand at ordinary temperatures for three weeks.
The product, which was slightly brownish, was then poured on ice,
and the solution several times extracted with pure ether. On drying
this ethereal solution over calcic chloride and evaporating, an almost
colourless oil was left which did not crystallise, even after standing
for several days over sulphuric acid in a vacuum, and therefore was
not analysed. When treated with dilute ammonia, the greater part
dissolved, leaving a small quantity of a brownish-coloured oil.

When heated in a test-tube, carbonic anhydride was given off, and
the residue on fractioning went over almost entirely between 205°
and 215° as a colourless oil, which was without doubt ethyl phenyl
ketone, C;H,-CO-CH,°CH; (b. p. 210°).

The solution of the crude acid was coloured reddish-brown on the
addition of a drop of ferric chloride. It is therefore almost certain
that this oily acid consisted for the most part of methylbenzoyl-
acetic acid, rendered impure by the presence of oily decomposition
products which prevented it from crystallising.

Action of Phosphorous Pentachloride on Ethylic Methylbenzoylacetate.

a-Methyl-8-chlorocinnamic Acid.

On treating ethylic benzoylacetate with pentachloride of phos-
phorus, it is transformed (Trans., 1885, 256) into a chlorocinnamic
acid, CH; : CH : CCl: COOH, in which the chlorine-atom appears to
be in the a-position. From the method of formation, however,
8-chlorocinnamic acid should be formed according to the equation—

I. C,H;-CO-CH,-COOEt + 2PCl, = C,H,°CCl,-CH,*COCI
+ EtCl + 2POC\,.

II. C,Hs-CCl,CH,-COCl + H,O = C,H;-CCl: CH-COOH + 2HCI1.

{t appeared therefore interesting to follow up this reaction, and in the
first place to examine the action of pentachloride of phosphorus on a
substituted ethylic benzoylacetate.

For this purpose, 10 grams of ethylic methylbenzoylacetate were
dissolved in 50 grams of phosphorous oxychloride, and then 25 grams
of phosphorous pentachloride were slowly added, the whole being
well cooled with ice during the operation. After standing for about
an hour, the mixture was finally heated a short time on a water-bath
and then poured on ice. When all the oxychloride had been decom-
posed, a yellowish oil was left, which was separated from the acid
liquid by extraction with ether. The ethereal solution was shaken

158 PERKIN AND CALMAN: BENZOYLACETIC ACID —

with dilute sodic hydrate to remove any acids which had been formed,
dried over calcic chloride and the ether distilled off.

In this way a yellowish oil was obtained which was further purified
by fractioning under reduced pressure (300 mm.). Almost the whole
of it went over in the first distillation between 225° and 240°, and
after repeated fractioning an oil was obtained boiling constantly at
230—233° (300 mm.). This gave the following numbers on analysis :—

I. 0°1996 gram substance gave 0°4700 gram CO, and 0°1076 gram
H,0.
II. 0°1536 gram substance gave 0°0952 gram AgCl.
Found.

sO Theory.
I. Il. C,H;-CCl ; C(CH;)-COOC,H;.

C .... 6422 — per cent. 64°14 per cent.
H.... 9599 — " 5°79 -
Cl.... = 15°34 0 15°81 -

This substance therefore is ethylic a-methyl-f-chlorocinnamate,
formed according to the following equation :—

C,.H,;-CO-CH(CH;)-COOC.H; = PCI; = C,H;'CCl : C(CH;)-COOC.H;
+ HCl + POC).

On examining the alkaline solution obtained by extracting the
crude product of the action of phosphorus pentachloride on ethylic
methylbenzoylacetate, it was found that when this was acidified
with dilute sulphuric acid, a quantity of an organic acid was precipi-
tated in white flakes. These when collected, dried on a porous plate,
and crystallised several times from boiling light petroleum, were ob-
tained in the form of colourless needles, which gave the following
results on analysis :—

I. 0:1890 gram substance gave 0°4213 gram CO, and 0°086 gram
H,0.
II. 0°2413 gram substance gave 0°1555 gram Ag(Cl.

Found.

Theory.
I ; C,H;-CCl: C(CH;)-COOH.

C .... 61°07 per cent. 60°79 per cent.
BZ... 68 — - 5°06 "
GG... = 18:07 - 17°95 -

This acid is therefore a-methyl-f-chlorocinnamic acid. It melts
at 116°, and can be distilled in small quantities almost without de-
composition. It dissolves easily in alcohol, ether, chloroform, benzene,
and acetic acid, more sparingly in cold light petroleum. When

AND SOME OF ITS DERIVATIVES. 159

treated with concentrated nitric acid, crystalline nitro-derivatives are
formed, which, however, were not further examined.

The silver salt of this acid was prepared by adding silver nitrate
to a neutral solution of the ammonic salt. After collecting and
washing with water, it was obtained as a white mass, which when
dried over sulphuric acid in a vacuum gave the following numbers on
analysis :—

0'3175 gram substance gave 0°1482 gram AgCl.

Theory.
Found. C,;H;-CCl : C(CH;)-COOAg.
35°20 per cent. 35°64 per cent.

On heating this salt, it is decomposed, a colourless oil being given
off which is probably chloromethylstyrene. In this chloromethyl-
cinnamic acid, it is obvious that the chlorine-atom must be in the
B-position, the «-position being occupied by the methyl-group.

a-Methyl-B-phenylhydroxypropionic Acid.

Ethylic benzoylacetate is easily reduced when treated with sodium
amalgam, f-phenyllactic acid being formed (Trans., 1885, 253),
thus—

C;H,;-CO-CH,.COOC,H; + H, + H.O = C,H;;CH(OH)-CH,,COOH
+ C.H;-OH.

In order to determine whether this reaction is a general one, it was
determined to try the action of reducing agents on a substituted
ethylic benzoylacetate. For this purpose, pure ethylic methyl-
benzoylacetate was dissolved in dilute alcohol, and sodium amalgam
slowly added, the whole being cooled from time to time with cold
water. After 24 hours, the alkaline liquid was separated from the
mercury, evaporated on a water-bath until all the alcohol had been
driven off, acidified with dilute sulphuric acid, and extracted six times
with ether. On drying over calcic chloride and distilling off the
ether, a thick oil was left, which did not crystallise even after re-
maining for some hours over sulphuric acid ina vacuum. In order
to purify this substance, it was dissolved in a slight excess of baryta-
water, the excess of baryta precipitated by carbonic anhydride and
the whole filtered. The filtrate was then acidified with dilute hydro-
chloric acid and again extracted with ether. On distilling off the
ether, a colourless oil remained, which solidified on standing. The
crystalline mass was first spread out on a porous plate to remove
traces of oily matter, and then dissolved in a little boiling benzene.
On cooling, the acid crystallised out in colourless needles which gave
the following numbers on analysis :—

160 PERKIN AND CALMAN: BENZOYLACETIC ACID

I. 0°1920 gram substance gave 0°4695 gram CO, and 0°1175 gram
H,0.
II. 0°2671 gram substance gave 0°6488 gram CO, and 0°1666 gram
H,0.
Found.

——_{ Theory.
I. II. C,;H;-CH(OH)-CH(CH,)-COOH.

C .... 6635 66°24 per cent. 66°67 per cent.
H.... 6°80 693 Csi, 667,
O.... 2685 2683 ,, 2667 ~~,

a-Methyl-8-phenylhydroxypropionic acid when quite pure melts at
124—125°. It is very easily soluble in alcohol, ether, acetone, and
hot water, sparingly soluble in chloroform, bisulphide of carbon, and
light petroleum. It crystallises beautifully from boiling water in
nodular masses, and from a solution of equal parts of light petroleum
and benzene in flat needles. If the free acid is heated in a test-tube,
water is given off and an oil distils over, which on cooling for the
most part solidifies, and smells strongly of benzaldehyde.

When mixed with concentrated sulphuric acid, it becomes yellowish,
and on warming it dissolves forming a reddish-brown solution.

The ammonium salt of a-methyl-8-phenylhydroxypropionic acid
crystallises from water in beautiful feathery needles. The silver salt
is easily obtained on adding silver nitrate to a neutral solution of
the ammonium salt as a heavy white crystalline precipitate, which
gave the following numbers on analysis :—

0°2773 gram substance gave 0°1036 gram silver.

Theory.
Found. C,H;-CH(OH)-CH(CH;)-COOAg.

37°36 per cent. 37°63 per cent.

This silver salt, when heated, was rapidly decomposed into silver, and
a colourless oil, which smelt strongly of benzaldehyde.

Ethylic Propylbenzoylacetate.

This ethereal salt was prepared by heating a mixture of ethylic
benzoylacetate (50 grams) with sodic ethylate (Na = 6 grams) and an
excess of propylic iodide (50 grams) for four hours at 100° in a soda-
water bottle.

The product was isolated in the same way as the ethylic methyl-
benzoylacetate and purified by fractioning under diminished pressure
(300 mm.). In the first distillation, almost all went over between
240° and 260°, and after repeated refractioning the ethylic propyl-

AND SOME OF ITS DERIVATIVES. 161

benzoylacetate was obtained pure as a colourless oil, boiling con-
stantly between 250° and 252° (300 mm.). On combustion it gave
the following numbers :—

0°3358 gram substance gave 0°8886 gram CQ, and 0°2423 gram

Theory.
Found. C,H;-CO-CH(C,H;)-COOC,H;.
72°15 per cent. 71°79 per cent.
7°69 1”
20°52 7

Ethylic propylbenzoylacetate is a colourless oil of aromatic odour,
which does not solidify at 0°. It is slightly soluble in caustic potash.
When treated with sodic ethylate, it gives a sodium-derivative. Its
alcoholic solution gives no coloration with ferric chloride.

When quite pure, it was found to have the following boiling
points :—

250—252° (300 mm.)
238—239 (225 ,, )

Butyl Phenyl Ketone.

In order to obtain this ketone, ethylic propylbenzoylacetate was
heated in a flask connected with a reflux condenser with dilute
alcoholic potash for six hours, at the end of which time decomposition
was complete. On adding water, an oil was precipitated, which was
separated from the alkaline solution by extraction with ether. The
ethereal solution, after being well washed with water, was dried over
potassic carbonate, and the ether distilled off.

In this way a dark-brownish oil was obtained, which after repeated
fractioning boiled constantly between 236° and 238° (720 mm.), and
on analysis gave numbers, showing it to be butyl phenyl ketone :—

I. 0°1158 gram substance gave 0°3441 gram CO, and 0:0957 gram
H,0.
II. 0°1805 gram substance gave 0°5373 gram CO, and 0°1528 gram
H,0.
Found.

-—-o_eo Theory.
4 II. C;H,;-CO-CH,-CH,-CH.-CH3.

81°05 81:19 per cent. 81:48 per cent.
H.... 919 928 Sis, 864 ,,
O.... 975 953s, 988 —éi,,

This ketone is formed from ethylic propylbenzoylacetate by hydro-
lysis and splitting off of carbonic anhydride according to the following
equation :—

162 PERKIN AND CALMAN: BENZOYLACETIC ACID

C,H,-CO-CH(C;H,)-COOC,H; aa 2KOH — C,H,;'CO:CH,°C,H, +
K,CO, + C.H;°OH.

It is a colourless oil of agreeable aromatic odour, and boils at 236—
238° (720 mm.) without decomposition.

Action of Phosphorus Pentachloride on Ethylic Propylbenzoylacetate.

It was next thought interesting to try the action of phosphorus
pentachloride on ethylic propylbenzoylacetate. 10 grams of the sub-
stance were dissolved in 35 grams of oxychloride of phosphorus,
18 grams of phosphorus pentachloride were slowly added, and the
whole allowed to remain until the principal reaction was over. The
mass was then heated on a water-bath for half an hour, poured into
ice, and from time to time stirred round until the oxychloride of
phosphorus was completely decomposed. After extracting with ether,
the ethereal solution was treated with dilute sodic hydrate, then dried
over calcic chloride and the ether distilled off. In this way, a heavy
yellow oil was obtained which after repeated fractioning in a vacuum
was obtained pure as a colourless oil boiling between 247° and 249°
(300 mm.).

On analysis the following numbers were obtained :—

I, 0°2251 gram substance gave 0°5480 gram CO, and 0°1438 gram
H,0.
II. 0°2085 gram substance gave 0°5075 gram CO, and 0°1228 gram
H,O.
III. 0°1330 gram substance gave 0°0755 gram Ag(Cl.

Found.
- ail ~ Theory.
L . Ut. C,H,-CC1: C(C,H,)-CO00,H,
C.... 6642 6645 — per cent. 66°67 per cent.
H.... 709 656 — ” 6°75 i
Cl.... — — 1407 - 13°89 ”

This substance was therefore ethylic «-propyl-f-chlorocinnamate,
formed according to the equation—

C.H;CO-CH(C;H,)-COOC,H + PCI; =
C.HsCCl1: C(C;H,)-COOC.H; + POCI, + HCl.

The alkaline solution, formed by treating the ethereal solution of
the crude product of the action of phosphorus pentachloride on
ethylic propylbenzoylacetate with sodic hydrate was next examined.
On acidifying with dilute sulphuric acid, yellow flakes were precipi-
tated, which were collected and well washed with water. They were

AND SOME OF ITS DERIVATIVES. 163

then dried on a porous plate and purified by several recrystallisations
from light petroleum.
The following analytical results showed that the compound is
a-propy]8-chlorocinnamic acid :—
I. 0°2030 gram substance gave 0°4738 gram CO, and 0°1022 gram
H,0.
II. 0°2246 gram substance gave 0°5289 gram CO, and 0°1156 gram
H,0.
III. 0°2839 gram substance gave 0°1789 gram AgCl.

Found.
cC- A —~ Theory.
L. I. I. C,H;-CCl: C(C,H;)-COOH.
63°66 6423 — per cent. 64°29 per cent.
562 572 — ” 5°80 ”
— — 15°39 ” 15°63 9”

a-Propyl-8-chlorocinnamic acid melts at 121°, and sublimes in long
colourless needles when carefully heated. It is easily soluble in alcohol,
ether, chloroform, benzene, and acetic acid, more sparingly in cold light
petroleum. By slowly cooling its solution in boiling light petroleum, it
is obtained in beautiful prisms, which Professor Haushofer was kind

enough to measure. He gave us the following description :—

Crystalline System. Triclinic.
a:b:¢: = 07967: 1 : 0°7403
a = 122°
b = 106
c= 69

coPco (010) = b
coPco (100) = a
OP (001) =c
oP, (110) = p
'‘P'oo (101) = r

It forms tabular shaped twinned crystals on the surface b, with
which the two surfaces a form a projecting angle in front, and a
re-entering angle behind of 151° 22' (measurement 151° 22—42’).
The surfaces c form a projecting angle above and below a re-entering

164 PERKIN AND CALMAN: BENZOYLACETIC ACID

angle of 121° 19’. The re-entering angle between the surfaces ¢
could not be measured, as the crystals were irregular at the ends.
The surfaces } are generally curved, and not adapted for measure-
ment, the edges in most of the crystals are broken up, but occasionally
they are fully developed, and brilliant as the diamond. No single

crystals were noticed.
The surfaces p are usually wanting in the twins, often r also.

Measured. Calculated.
*75° Al’
*96 44
*119 16
#133 «52
*140 39
101 43 101° 56’
121 19 121 19 (in the twins).

In convergent polarised light two systems of rings are observed on
the surface b of the twins, the optical axes of which belong to both
the twins.

The silver salt of the acid, prepared by precipitating a neutral
solution of the ammonium salt with silver nitrate, is a thick white

amorphous mass, almost insoluble in water. The analysis gave the
following numbers :—
0°4945 gram substance gave 0°2165 gram AgCl.

Theory.
Found. C,H;-CCl : C(C,;H;)-COOAg.

32°95 per cent. 32-25 per cent.

Ethylic Isopropylbenzoylacetate.

In order to obtain this substance, 50 grams of ethylic benzoylacetate
were heated with 6 grams of sodium and an excess of isopropylic iodide
for six hours at 100°. Water was then added, and the product
isolated in the usual manner. On distilling the crude substance,
almost all went over at 225—245° (225 mm.), and on repeated refrac-
tioning an oil was obtained boiling constantly at 236—237° (225 mm.),
which gave the following numbers on analysis :—

0°2760 gram substance gave 0°7230 gram CO, and 0°188 gram H,0.

Theory.
Found. C,H,-CO.CH[C(CH,).]-COO0,H,.
71°45 per cent. 71°79 per cent.
7°61 - 7°70 ”
20°51

AND SOME OF ITS DERIVATIVES. 165

Ethylic isopropylbenzoylacetate is an agreeably smelling oil. Its
boiling point is about 2° lower than that of the normal propylic
derivative previously described.

Isobutyl Phenyl Ketone.

This ketone was prepared by digesting ethylic isopropylbenzoyl-
acetate with dilute alcoholic potash for about six hours. The brownish-
coloured product was then diluted with water and several times
extracted with ether. After distilling off the ether, a dark-coloured
oil was left, which was easily obtained colourless by distillation. On
repeatedly refractioning it, it boiled constantly at 227—228° (720 mm.),
and gave on analysis the following numbers agreeing with the
formula CsH;-CO-CH,CH(CHs), :—

01605 gram substance gave 0°4775 gram CO, and 0°1315 gram

Found. Theory. C,,H,,0.
81:14 per cent. 81°48 per cent.
” 864
" 9°87 -

Isobutyl phenyl ketone is a colourless oil, of an agreeable aromatic
odour.

This ketone has already been prepared by Popow (Annalen, 162,
153) by distilling a mixture of benzoate and valerate of calcium.
Popow gives the boiling point as 225—226°.

Isobutyl phenyl ketone does not combine with hydrogen sodium
sulphite. On oxidation with chromic acid, benzoic, isobutyric, and
acetic acids are formed (Popow).

Ethylic Isobutylbenzoy lacetate.

This substance was prepared in exactly the same way as the
isopropyl-derivative by heating ethylic benzoylacetate (1 mol.) with
sodic ethylate (1 mol.) and an excess of isobutylic iodide at 100° for six
hours. On fractioning the product, almost the whole of it went over
between 230° and 260° (225 mm.) in the first distillation. Pure
ethylic isobutylbenzoylacetate boils at 246—247° (225 mm.).

The following results were obtained on analysis :—

0'2257 gram substance gave 0°5960 gram CO, and 0°1620 gram

Theory.
Found. C,H,-CO-CHICH,-CH(CH;),]-COOC,H,.
72°01 per cent. 72°58 per cent.
7°97 ” 8°06 ”
20°02 19°35 om

166 PERKIN AND CALMAN: BENZOYLACETIC ACID

This ethereal salt has much the same properties as the isopropyl-
derivative. The boiling point is about 10° higher.

Isoamyl Phenyl Ketone.

Ethylic isobutylbenzoylacetate is easily split up on heating it with
dilute alcoholic potash into alcohol, carbonic anhydride, and isoamyl]
phenyl ketone, thus :—

C,H;-CO-CH[CH,-CH(CH;).|*COOC,H; + 2KOH
= C,;H;-CO-CH:.CH,*[ CH(CHs)2 | + K,CO, + C.H;°OH.

The product of this reaction was isolated in the usual manner, and
purified by fractioning. Isoamyl phenyl ketone is an aromatic
smelling oil, boiling at 240—241° (720 mm.). Analysis :—

0°1870 gram substance gave 0°561 gram CO, and 0°156 gram H,0.

Theory.
Found. C,H,-CO-CH,-CH,-CH(CH,)-.
81°82 per cent. 81°82 per cent.
” 9°09 ”
” 909s,

In Part III of this research (Trans., 1885, 265—266) an acid,
C,,H;,0;, was described, which was obtained in small quantity from
ethylic dibenzoylsuccinate by prolonged boiling with dilute sulphuric
acid, and which from its method of formation was thought to be simi-
larly constituted to carbopyrotritartaric acid.

This appeared the more probable from a small experiment which
was tried at the time (loc. cit., p. 271) on the action of concentrated
sulphuric sulphuric acid on ethylic dibenzoylsuccinate ; a compound
was produced which on hydrolysis yielded the acid C,,H,,0O; quanti-
tatively, and was therefore probably its ethereal salt. In order to
determine the nature of this decomposition, the experiment has since
been carried out on a larger scale, in the following way:—

Finely-powdered ethylic dibenzoylsuccinate was shaken with about
10 times its weight of pure concentrated sulphuric acid, in which it
soon dissolved, forming a colourless solution. This was allowed to
remain for 12 hours, and then poured on to ice, when a thick almost
colourless oil was precipitated, and was separated from the acid liquid
by extracting it with ether. The ethereal solution was then well
washed with water, and dried over calcic chloride. On distilling off
the ether, an almost colourless oil remained, which instantly became
solid on being touched with a glass rod. The mass was then dissolved
in hot alcohol, from which it crystallised on cooling in beautiful canary-
yellow prisms.

The analysis gave the following numbers :—

AND SOME OF ITS DERIVATIVES. 167

I. 0°1698 gram substance gave 0°4505 gram CO, and 0°0900 gram
H,0.
IJ. 0°2250 gram substance gave 0°5974 gram CO, and 0:1134 gram
H,0.
Found.
\—' Theory.
I. CysH 9035.
72°41 per cent. 72°53 per cent.
5°60 ” 5°89 ”
21°99 “ 21°98 ”
This substance contains 1 mol. H,O less than ethylic benzoyl-
succinate, and appears to be formed according to the following equa-

tions :—

C.H;.CO CO-C,H, C,H,.C(OH) C(OH)-C,H,.
|| |
COOEt-CH—CH:COOEt COOEt-C-——C:COOEt

Ethylic dibenzoylsuccinate. Labile form of the same.
a
C,H;°C C:C,H;
= i | || + H,0.
COOEt-C C-OOEt
Ethylic diphenylfurfurandicarboxylate (1, 4, 2, 3).

Ethylic diphenylfurfurandicarboxylate [1, 4, 2, 3] melts at 86°,
and is characterised particularly by the readiness with which it
crystallises. It is easily soluble in chloroform, benzene, ether, alcohol,
and light petroleum, and can be obtained in beautiful almost colour-
less prisms, by allowing the ethereal solution to evaporate slowly.
Prof. Haushofer was kind enough to measure these. He gave me the
following description of these crystals :—

Crystalline System. Rhombie.
a:b:c = 0°9005 : 1 : 0°6309

168 PERKIN AND CALMAN: BENZOYLACETIC ACID

Pale sulphur-yellow transparent crystals of the combinations
P(111) = 0, coP4(210) = p, coPcc(010) = b. Short prismatic in the
vertical axis, usually tabular through development of the surfaces b.

Measured. Calculated.

0:0 = (111) (111) = *86° 38’ — — (Basal angle).

o:0 = (111) (111) = *118 42 — — (Macrodiagonal. Polar
angle).

:o = (111) (111) = 125 10 125° 20' (Brachydiagonal. Polar

angle).

»: p’= (210) (210) = 181 382 131 35

o:p = (111) (210) = 180 37 130 47

o:p'= (111) (210)= 106 38 106 50

The base of the plane the optical axis. In convergent polarised light
the interference image of an axis (excentric in the direction of the
angle pb) appears on the surface p.

When heated in small quantities, this ethereal salt distils almost
without decomposition. Although the solution in chloroform does
not decolorise bromine at ordinary temperatures, it does so easily on
warming, hydrobromic acid being liberated. When treated with
nitric acid, beautifully crystalline nitro-derivatives are formed.

Ethylic diphenylfurfurandicarboxylate is not acted on by phosphorus
pentachloride at ordinary temperatures. It dissolves in concentrated
sulphuric acid, forming a colourless solution, which on heating
becomes first brownish and then of a beautiful violet colour.

Diphenylfurfurandicarborylic Acid [1, 4, 2, 3].

When ethylic dibenzoylsuccinate is boiled with dilute sulphuric
acid, a small quantity of an acid is formed, which from previous
analysis (Trans., 1885, 267) was found to have the formula C,.H,.0;.

Owing to the small yield obtained by this reaction, it was only
superficially examined. The same acid can, however, be obtained
quantitatively by the hydrolysis of ethylic diphenylfurfurandi-
carboxylate with alcoholic potash. It is only necessary to digest
the ethereal salt 15 minutes with an excess of alcoholic potash, to
completely saponify it. At the end of the reaction, water is added,
and the whole evaporated until all the alcohol has been driven off.
On acidifying with dilute sulphuric acid, the acid is thrown down in
white flakes. After collecting and recrystallising it from acetic acid
(70 per cent.), it is easily obtained pure in colourless needles, melting
at 238°.

The following analyses confirm the formula previously assigned to
this acid.

AND SOME OF ITS DERIVATIVES. 169

T. 01762 gram substance gave 0°4523 gram CO, and 0°0637 gram
H,0.

II. 0°1848 gram substance gave 0°4735 gram CO, and 0:0726 gram
H,0.

Found.
“e Theory.
I. II. C5105.
C.... WG 69°89 per cent. 70°13 per cent.
H.... 401 436 =O, 389 Si,
O.... 2598 25°75 am 25°97 "

This compound is therefore diphenylfurfurandicarboxylic acid,

O
\
f C-C,Hs
i
COOH-C——C-COOH
It melts at 238° to a colourless liquid, which when strongly heated
splits up into the compound C,,H,O, and water. This was described

in Part III of this research (Trans., 1885, 268), and appears to be the
anhydride of diphenylfurfurandicarboxylic acid,*

CO-C : C(C.Hs)

1 oneal

According to this interpretation, the re-formation of the acid
CsH,.0; by treating this body C,,H,O, with alcoholic potash can be
easily understood.

The salts of diphenylfurfurandicarboxylic acid were described in
Part III of this research (Trans., 1885, 268), and the analyses of the
silver and calcium salts given. Owing to the small quantity of acid
which could be prepared at that time, no satisfactory results were ob-
tained by the analysis of these salts. The experiments were therefore
repeated with larger quantities of material. The silver salt was pre-
pared by fractional precipitation of the ammonium salt with silver
nitrate, and analysed with the following results :—

I. 0°2235 gram substance gave 0°3380 gram CO,, 0°0520 gram
H,0, and 0°0925 gram Ag.
IT. 0°2010 gram substance gave 0°3060 gram CO,, 0°040 gram
H,0, and 0°0833 gram Ag.
IIT. 0°2750 gram substance gave 0°1130 gram Ag.

* This is also the opinion of Knorr (Ber., 17, 2869).
VOL. XLIX. N

PERKIN AND CALMAN: BENZOYLACETIC ACID

Found. Old analyses.

_
II. Ill.
4151 — pe.
221 — =,
41°44 41:09 , 40°52
1484 — ,, 16°50

Theory. C©,g,H,0;Ago.
41°44 per cent.

Diphenylfurfurandicarboxylic acid is therefore a simple bibasic
acid forming salts of the formula C,,H»O;M2.

From the method of formation of this acid it would at first sight
appear probable that its constitution was exactly similar to that of
carbopyrotritartaric acid, which Knorr has shown to be produced
by the action of concentrated sulphuric acid on ethylic diacetylsuc-
cinnate :

J Fas
C,H,C C-C,H; CH,C C-CH;

Id lI
COOH:C C-COOH COOH:-C C.COOH
Diphenylfucfurandicarboxylic Dimethylfurfurandicarboxylic acid
acid. (carbopyrotritartaric acid).

There is, however, one great difference between these acids; carbo-
pyrotartaric acid is easily decomposed on heating above its melting
point into pyrotritartaric acid and carbonic anhydride,

C,H,0; — CO, = C;H;0;,

whereas diphenylfurfurandicarboxylic acid is simply converted into its
anhydride, as described above, the monocarboxylic acid corresponding
with pyrotritartaric acid not having as yet been obtained.

Lately Fittig (Ber., 18, 3410) has made it doubtful whether carbo-
pyrotritartaric and pyrotritartaric acids really are furfurane-deriva-
tives.

In his researches on the condensation of ethylic acetoacetate with
succinic acid, he obtained an acid isomeric with carbopyrotritartaric
acid, according to the equation—

CH,°CO-CH,-COOC,H; + COOH:CH,CH,-COOH
= C,H.O; + H,O + C.H,°OH.

Methroniec acid.

AND SOME OF ITS DERIVATIVES. 171

which at a temperature of 200—240° is converted quantitatively int»
pyrotritartaric acid and carbonic anhydride.

In this way, it is difficult to conceive how pyrotritartaric acid can
contain two methyl groups, and Fittig therefore supposes that in the
formation of ethylic carbopyrotritartrate from ethylic diacetylsuccinate
condensation takes place in a way analogous to the formation of
mesityl oxide from acetone.

CH,;°CO-CH:-COOC,H,; CH;°C CH-COOC,H;
l

| | |
CH-COOC.H; CH CH-COOC,H;
\0/

Ethylic diacetylsuccinate. Ethylic carbopyrotritartrate.

Pyrotritartaric acid would then have the constitution—

CH,-C-——-CH:COOH

| |
CH CH,
\co%

It is therefore probable that the acids obtained from ethylic diacetyl-
succinate and dibenzoylsuccinate by the same reaction, 7.e., by the
action of dehydrating agents, are of an entirely different structure,
the former being derivatives of pentamethylene, the latter of fur-
furane. From the experiments described above, it is scarcely possible
that the acid C,,H,.O; can be a lactonic acid, as Fittig believes
(Ber., 18, 3411).

In conclusion, we append for convenience sake tables of the boiling
points of the homologues of ethylic benzoylacetate obtained in the
course of this research, and of the homologues of acetophenone pre-
pared by the hydrolysis of these ethereal salts.

Boiling points of the Ethylic Salts of Benzoylacetic Acid and its
Homologues under 225 mm. pressure.

Ethylic benzoylacetate 220—225°

methyl benzoylacetate 226—227

ethyl benzoylacetate 231—232

propyl benzoylacetate 238—239

isopropyl benzoylacetate...... 236—237

allyl benzoylacetate 240—241

isobutyl benzoylacetate 246—247
Methylic benzylbenzoylacetate 250—255 (?)

(50 mm. pressure).
n 2

BROWN: THE CHEMICAL ACTION OF

Boiling point under
720 mm. pressure.

Acetophenone

Ethyl phenyl ketone

Propyl phenyl ketone 220—222

Allylacetophenone 235—238

Butyl phenyl ketone 236—238
227—228
229---231
240-—241

Further experiments on ethylic benzoylacetate, and more especially
on the behaviour of ethylic dibenzoylsuccinate with different reagents,
are in progress.

XIX.—The Chemical Action of Pure Cultivations of Bacterium Aceti.
By A. J. Brown.

Tae following paper contains an account of experiments made for
the purpose of investigating some of the chemical actions of a pure
cultivation of Bacterium aceti.

Pasteur, in his “ Mémoire sur la Fermentation Acétique,” published
in 1864, was the first to show that acetic fermentation is caused by a
living micro-organism, which he named Mycoderma aceti. He further
shows that the acetic acid formed during the fermentation is produced
by the oxidation of ethylic alcohol by the oxygen in the air, this action
being brought about by the ferment. Moreover, when the fermenta-
tion is weakened in certain ways, aldehyde is present, and if the fer-
mentation is allowed to go on after the whole of the alcohol is
oxidised, the acetic acid already formed becomes further oxidised
to CO, and H,O. Pasteur also believes that succinic acid in small
quantity is always formed as a bye-product during the acetic fermen-
tation.

E. Wurm (Dingl. polyt. J., 235, 225) investigated the acetic fer-
mentation, and confirms Pasteur’s statement that the formation of
vinegar from alcoholic liquids is due to an organised ferment.

In the Comptes rendus of July, 1880, p. 236, Boutroux describes an
action of the acefic ferment on glucose, but this paper will be more
fully referred to later on.

nt —_e hte eee ~~ hom mae ar ~~ aa.

PURE CULTIVATIONS OF BACTERIUM ACETI. 173

Cohn (Biol. d. Pflanzen, 2, 173), and more particularly Hansen
(Meddelser fra Carlsherg-Laboratoriet), have investigated the acetic
ferment (B. aceti) morphologically, but do not appear to have studied
its chemical action.

Before describing my experiments, I think it advisable in a research
of this nature to describe the method of working, and the means
used for the purification of the ferment.

As B. aceti is strictly an aérobic ferment, and consequently as my
experiments required free access of air, the method of working with
flasks or test-tubes plugged with sterilised cotton-wool was found
very suitable. All the vessels and cotton-wool used, and also the
various liquids in which the ferment was grown, were sterilised with
all the precautions advised by Klein (“‘ Micro-organisms and Disease,”
Practitioner, 1884). No solution was considered sierilised and fit for
use unless it had kept quite free from all forms of living organisms
for at least a week at a temperature of 28°. Inoculation from one
vessel to another was effected by means of a capillary glass tube
freshly drawn out and pointed at one end, so that it could be forced
through the cotton-wool plugs in the necks of the vessels, and each
tube was not used more than once. The cotton-wool plugs of the
vessels were kept covered with sterilised paper, and, before inoculating
from one vessel to another, the surface of the plug was singed by a
gas flame.

I found the most suitable method for obtaining a pure culture of
B. aceti was a combination of Kleb’s “ Fractional” and v. Nigeli’s
“Thilution”” methods. Inoculation in gelatin media did not give
such satisfactory results, because B. aceti being so strictly aérobic it
only grows on the immediate surface of the gelatin, and in such an
attenuated film that it is difficult to detect it in its early stage of
growth before it has spread over a considerable surface.

The B. aceti used in my experiments was originally taken from the
surface of a beer that had been kept exposed to the air in a warm
place until the acetic fermentation developed. This ferment was
inoculated into the first of a series of 10 test-tubes, containing a
sterilised solution of 2 per cent. ethylic alcohol in yeast water ;* the
inoculations being made from one tube to another as soon as a visible
growth was observed (generally from 36 to 48 hours). Finally, the
culture was further tested by the “dilution” method, and showed its
purity by always producing a growth of B. aceti (or no growth at all)
in the tubes into which the diluted ferment was inoculated. Further,
if the reactions of this ferment, which I describe further on, are
considered, the comparative simplicity of the action in some cases,

* Used by Pasteur, and prepared by boiling 7 per cent. yeast, pressed as dry as
possible, in water, and filtering until quite bright.

174 BROWN: THE CHEMICAL ACTION OF

and the want of action altogether in others, will leave no doubt that
my experiments were made with a pure cultivation of acetic ferment.
B. aceti has so many different forms of growth, that I do not consider
the microscope alone gives much trustworthy aid in detecting foreign
ferments mixed with it.

The mode of growth of pure B. aceti in alcoholic or sugar solutions, is
in the form of a rather greasy pellicle covering the surface of the liquid,
and inclined in its young stage of growth to creep up the moist sides
of the containing vessel.

This pellicle varies in thickness from an almost invisible film on the
surface of such a mixture as dextrose and Pasteur’s solution, to the
thickuess of stout paper on diluted claret. Slight agitation easily
breaks the surface of this pellicle, and, after its somewhat greasy
surface is wet, it sinks to the bottom of the liquid, only, however, for
another growth to form again on the surface, if the liquid still
contains the proper nourishment for it. The liquid below the pellicle
is usually turbid with isolated cells of the ferment, and, after a culti-
vation has been allowed to remain undisturbed for some weeks, a con-
siderable deposit of the growth is found at the bottom. B. aceti is,
strictly speaking, an aérobic ferment; if inoculated into solutions kept
free from oxygen, it will not increase or produce any chemical changes
in the liquid, but, nevertheless, is capable of living under these cir-
cumstances for a long time. I have kept an inoculated solution out
of contact with air for six months without its showing any signs of
growth, but on admitting filtered air a strong growth started in a
few days.

Hitherto, two species of bacteria said to have the power of oxi-
dising alcohol to acetic acid have been described, viz., the ordinary
B. aceti with which we are now concerned, and the B. Pastewrianum
of Hansen. There is, however, at least one other distinct species,
about which I hope to say something in another communication.
Under these circumstances it seems advisable to describe shortly the
morphology of the species with which I have worked (viz., B. aceti).

B. aceti in its normal state, freshly growing as a pellicle on the
surface of a liquid, appears under the microscope as a mass of cells
about 2u* in length, and slightly contracted in the middle, giving
them a sort of figure of 8 appearance. These cells are united into
chains of variable length, which are easily broken up by pressure of
the cover-glass. Frequently the cells are quite divided in the middle,
thus producing strings of micrococcas-like forms; both forms being
sometimes found in the same chain. The above two forms are those

* No exact measurements of B. aceti can be given, as all the forms differ very
much amongst themselves in size, varying with the different culture fluids, amount
of acid, &e.

PURE CULTIVATIONS OF BACTERIUM ACETI. 175

usually present when the ferment is growing vigorously at the surface
of a liquid. But in the liquid below the surface film, and on the
bottom of the containing vessel, abnormal forms are often found
differing very much from the ordinary surface growth; this is more
especially the case in old cultivations. These forms often attain the
length of 10—1l5,, or even more; in some cases their form is that of
leptothrix threads of even thickness throughout their length ; in others,
the long cells are swollen out in two or three places along their length,
giving them a most irregular appearance. These cells are generally
of a dark-grey colour. At their ends a short chain of short rods or
micrococci is sometimes observed. The other forms most frequently
seen are short rods about 3 in length, and micrococci about ly in
diameter, floating freely in the culture liquid. My observations on the
morphology of B. aceti agree very well with the drawings given by
Hansen (Meddelser fra Carlsberg-Laboratoriet, 2, 1879), and by Zopf
(Die Spaltpilze, 1885, pp. 9 and 63).

The shorter rods and cells of B. aceti, when floating freely in culture
fluid, are motile, but I have not noticed this property in the large
abnormal forms.

When treated with dilute solution of iodine, B. aceti is stained
yellow. Hansen states that the only difference between this ferment
and B. Pasteurianum is that the latter is stained blue by similar
treatment.

Boiling with dilute caustic potash quickly disintegrates the pellicle
of B. aceti; also on treating it with strong sulphuric acid it is at
once broken up, and the further addition of iodine merely turns it
yellow.

. Oxidation of Ethylic Alcohol.

Klein (‘ Micro-organisms and Disease,” 1884), when describing
B. aceti, states that “‘ pure cultivations have not been made with it,
and before deciding whether it is the real cause of the acetic acid
fermentation, experiments with such pure cultures, 7.e., inoculations of
alcoholic fluids with it, are required.” For this reason, the first
experiment I describe concerns this well-known action.

A litre flask half full of a 5 per cent. solution of pure ethylic alcohol
in yeast water, and quite free from acid, was sterilised by the method
described above. It was afterwards kept for 10 days at a tempera-
ture of 28° to be quite sure of its sterility. This being ascertained,
the solution was inoculated with a trace of the pure cultivation of
B. aceti. The flask was then placed on a hot water tray connected
with a thermostat, and there kept at a constant temperature of 28°.
On the second day after inoculation, a fine film of ferment had com-

176 BROWN: THE CHEMICAL ACTION OF

menced to grow on the surface of the solution, which film increased
in substance slightly during the next few days. On the tenth day,
the flask was opened and found to smell strongly of acetic acid.
100 c.c. of the solution was distilled, pure water being frequently
added to the retort, until the whole of the volatile acid had passed
over. This volatile acid on titration gave a percentage of 1°021 acid
calculated as acetic acid on the original 100 c.c. of solution distilled.
The residue in the retort showed only 0-006 per cent. of non-volatile
acid. There has been formed, therefore, during the growth of B. aceti
in the original alcoholic solution, 1-021 per cent. of volatile acid and a
mere trace of non-volatile acid. In order to ascertain the nature of
the volatile acid, the rest of the solution was carefully distilled, and
the very acid distillate neutralised with an excess of pure baric car-
bonate. After filtration, the whole of the solution was evaporated to
dryness, and the residual salt dried at 130°. A weighed portion of
the salt was decomposed in a platinum crucible by sulphuric acid,
ignited, and the residual baric sulphate again weighed ; on calculating,
the following result was obtained :—

”

showing conclusively that acetic acid was the one and only volatile
acid formed in the above experiment with a pure cultivation of
B. aceti.

The trace of non-volatile acid formed in the experiment seemed to
answer to the tests for succinic acid, thus agreeing with Pasteur's
observations ; but the quantity found was too small for satisfactory
identification. Traces of a body resembling aldehyde are generally
found on distilling an acetic fermentation of B. aceti; this is more
especially the case when the ferment is growing with an insufficient
supply of oxygen.

Pasteur stated in his “Mémoire” that acetic acid is completely
burnt to carbonic acid and water by B. aceti, when there is no alcohol
in the solution. I confirmed this by the following experiment. A
solution of 0°75 per cent. acetic acid in yeast water, sterilised as usual,
was inoculated with B. aceti. The ferment grew with difficulty, and
only formed an extremely thin pellicle on the surface of the liquid.
After six weeks, the flask was opened and the acid determined. Only
0°25 per cent. of acid remained, showing that 0°5 per cent. of acid
had disappeared during the growth of the ferment. A second flask,
not inoculated, was found during the same time not to have lost
any appreciable amount of acid. If an alcoholic solution is allowed
to ferment until the whole of the alcohol is converted into acetic acid,
I find that the acid so formed is much more quickly decomposed by

177

PURE CULTIVATIONS OF BACTERIUM ACETI.

the ferment than in the experiment I have just described ; this most
probably is due to the strong pellicle of ferment formed during the
time the oxidation of the alcohol was going on.

Action of B. aceti on Propylic Alcohol.

Seeing that ethylic alcohol is so freely oxidised by B. aceti, it
appeared desirable to ascertain whether this action did not extend to
the other alcohols of the same series. I first experimented on normal
propylic alcohol, and prepared a solution containing 3 per cent. of
the pure alcohol in yeast water. After sterilising, &c., in the usual
way the solution was inoculated with B. aceti. For the first few days
the ferment grew very slowly, but later a fine pellicle was developed
on the surface. The flask was opened 14 days after inoculation, when
the liquid was found to have a strongly acid odour. A sample taken
for estimation of the total acid gave 1:20 per cent. calculated as
CH;COOH. The rest of the solution was distilled, and the acid
distillate neutralised with carbonate of barium, filtered and evaporated
to dryness. After drying at 130°, a weighed portion of the salt was
decomposed with sulphuric acid, ignited, and weighed. On calculation,
the following result was obtained :—

eeeerereere ee ee eee eeeeeee

oeeeeeee

showing that normal propylic alcohol is oxidised to propionic acid by
the action of B. aceti, just as ethylic alcohol is oxidised to acetic acid.
But a mere trace of non-volatile acid was found in the original
solution after distilling off the propionic acid.

Action of B. aceti on other Alcohols.

The next experiments were made with methylic alcohol. I found
this alcohol had to be purified by repeated distillation to get rid of all
traces of resinous matter before the ferment would grow freely in its
presence.

A 1 per cent. solution of this in yeast water was inoculated with
B. aceti. In a few days the ferment was growing freely. On opening
the flask after three weeks, the solution was found slightly alkaline.
(Before going further I had better add that B. aceti grows freely
in yeast water alone, the reaction becoming slightly alkaline. On
distilling this solution, ammonia or ammonium carbonate distils
over. The residue, after slightly acidifying with a few drops of
sulpharic acid, yields a very small quantity of volatile acids which
have a slight reducing action on silver nitrate and permanganate, and

178 BROWN: THE CHEMICAL ACTION OF

which are probably a mixture of butyric with a little formic acid.)
The alkaline distillate on examination contained only ammonia and
methylic alcohol. The residue in the flask, after slightly acidifying
with a few drops of sulphuric acid, was again distilled. A very small
quantity of volatile acid was obtained, which had a slight reducing
action on AgNO, and on HgCl,, but the very small quantity of
the acid and also its character is only what may be obtained from a
fermentation of yeast water alone.

Another experiment was made on a solution of yeast water con-
taining 1 per cent. of methylic alcohol and some calcic carbonate.
This fermentation was allowed to go on for four weeks, during which
time the ferment grew very freely. On opening the flask and dis-
tilling, after slightly acidifying with sulphuric acid, only 0:008 per
cent. volatile acid was found. From the above two experiments, it is
evident that B. aceti is unable to oxidise methylic alcohol to formic
acid, under circumstances in which it acts freely on ethylic and pro-
pylic alcohols. But as it was possible that methylic alcohol might
have been oxidised directly to carbonic acid and water in the experi-
ments quoted, a flask containing 500 c.c. of yeast water was sterilised,
and afterwards exactly 5 c.c. of methylic alcohol, sp. gr. 0°8151, was
added, and the solution inoculated with B. aceti. In order that the
alcohol in the solution should not be lost by diffusion of its vapour
through the cotton-wool plug, the mouth of the flask was closed by
an india-rubber stopper pierced by two tubes, so arranged that
filtered pure air could be drawn through the flask every second day,
the air that came from the flask being passed through a chloride of
calcium tube to arrest any methyl alcohol that might be with it.
The ferment in this experiment grew with great freedom, and the
fermentation was allowed to go on for four weeks and a half. On
opening the flask, the solution was carefully distilled until all the
alcohol had passed over, and the sp. gr. of the distillates was taken.
On comparing these weights with a table, I found 4°90 c.c. of alcohol
had distilled over, against the original 5°00 c.c. with which the experi-
ment had been started. These results agree as closely as could be
expected considering the method used, and leave no doubt that,
under the circumstances of my experiment, methylic alcohol is not
acted on in any way by B. aceti. Why this should be so with an
easily oxidisable liquid like methylic alcohol, when ethylic and pro-
pylic alcohols are acted on so readily, it is difficult to say. It can
hardly be because the products of the action of B. aceti on methylic
alcohol are poisonous to itself, for if so, why should it thrive so well
in the presence of the alcohol ?

I have made repeated attempts to oxidise isoprimary butylic

al uve — 7% —_— ~s

PURE OULTIVATIONS OF BACTERIUM ACETI. 179

alcohol (b. p. 108°) by means of B. aceti, but have been unsuccessful,
although the ferment will grow feebly in 0°5 per cent. solution of this
alcohol in yeast water.

I have not succeeded in making B. aceti grow at all in presence of
amylic alcohol (fermentation alcohol).

Action of B. aceti on the Carbohydrates.

In the Comptes rendus of March, 1878, p. 605, is a paper by
Boutroux entitled “Sur la fermentation lactique,” in which the
author describes, as a continuation of Pasteur’s work on the lactic
ferment (Ann. Chim. Phys., 1857), a purified cultivation of this
ferment, which grows as a pellicle on the surface of solutions
containing sugar and nitrogenous matter, and had the power of
converting this sugar into lactic acid. He further finds that this
ferment has the power of converting alcohol into acetic acid, and
concludes that the lactic and acetic ferment are one and the same
organism.

In the Comptes rendus of July, 1880, p. 236, another paper from
Boutroux appears correcting his former conclusions, and stating that
the acid formed by his ferment from sugar is not lactic but gluconic
acid. He gives analyses of the acid and its salts, and further states
that the ferment which forms gluconic acid is Mycoderma aceti,
and that he was in error in calling it lactic ferment.

It appeared desirable to repeat this experiment with my pure
cultivation of B. aceti. Boutroux in his experiments grew the
ferment in a solution of glucose in yeast water containing a quantity
of calcic carbonate to neutralise the acid when formed ;* but as there
is in this case a considerable quantity of organic matter from .the
yeast present which might possibly complicate the reaction, I
preferred to make my experiments with solutions containing merely
dextrose, and inorganic salts of known composition to serve as
food for the ferment. For this purpose, I prepared a flask containing
3 litres of a solution composed of 2 per cent. dextrose dissolved
in Pasteur’s mineral solution, and to this 10 grams of pure calcic
carbonate was added. This solution was sterilised, and after being
kept as usual was inoculated with B. aceti on August 2nd. The fer-
ment grew more slowly than in yeast water and dextrose, and the
pellicle formed on the surface was extremely thin. On September
10th, most of the calcic carbonate having been dissolved, the flask
was opened for examination.

* This acid fermentation of dextrose goes on freely in a solution without CaCO,

being added; but when about 0°4 per cent. acid (calculated as CH;-COOH) is
formed, further action is much retarded.

180 BROWN: THE CHEMICAL ACTION OF

A portion of the solution was distilled and examined for ethylic
alcohol, but none was found. The residue in the retort was then
rendered slightly acid with sulphuric acid, and the solution again
distilled. The distillate gave no acid reaction, showing the absence
of acetic and other volatile acids.

The bulk of the original solution was then filtered as clear as
possible, and evaporated slowly to a small bulk. To this solution, a
large excess of alcohol of 0°83 was added, which produced a bulky
precipitate of a brownish gummy nature. After allowing the whole
to stand 24 hours, the bright alcoholic solution was poured off, and
a small quantity of water added to the precipitate, which dissolved
it completely to a dark brown solution. After decolorising with
animal charcoal, a large excess of alcohol was again added; this
brought down the precipitate in nearly white flocks. The solution was
allowed to stand until quite bright, and then filtered. After washing
the precipitate thoroughly with alcohol and well draining, the salt
was again dissolved in water, and the solution heated to 100°.
Boiling alcohol was then added so as just to produce a slight
permanent milkiness, and the solution was put on one side to cool
slowly. After 48 hours, the bottom and sides of the beaker were
covered with beautifully white roundish concretions, evidently of a
crystalline nature. Under the microscope they were found to be
masses of minute acicular crystals.

After drying this salt at 100°, two estimations of the CaO in it
were made, giving—

No. I. 12°63 per cent. CaO.
No. Il. 12°68 __,, -

Calcic gluconate, according to Hertzfeld’s formula (Annalen, 220,
335), (C.Hi,0;),Ca + 1OH2, contains 12°50 per cent. CaQ.

0°6454 gram of this salt was burnt with lead chromate and
potassic dichromate. This combustion gave 0°3122 gram CO, and
0°7535 gram OH,. On calculation, the Ca being taken as the mean
of my two analyses, we get the following result :—

(C,H,,0,);Ca + 10H.
32°14
5°36
53°57

8:93

The above analyses leave no doubt that the salt with which we are
dealing is calcic gluconate.

This salt does not reduce Fehling’s solution, neither has its
solution any action on polarised light. It reduces silver nitrate with

PURE CULTIVATIONS OF BACTERIUM ACETI. 181

great ease, and also prevents the precipitation of ferric oxide by
ammonia. Its aqueous solution, on being freely exposed to the air for
several weeks, gradually deposits the crystalline form of the salt
again, unchanged. On heating the dry salt strongly, it intumesces
ina remarkable manner, and finally burns to a white ash. I have
prepared the free acid from the calcic salt; and the solution of the
free acid when evaporated in a vacuum leaves it as an uncrystallisable,
glassy, nearly colourless mass. Its aqueous solution is intensely acid
to the taste when concentrated, and on heating, even below 100°, it
begins to turn brown and decompose.

In my fermentation experiment with dextrose, calcic gluconate
appeared to be the only soluble calcic salt present. I therefore
examined the insoluble residue from the fermentation, but found it to
consist of calcic carbonate and cells of B. aceti alone. It therefore
appears that the sole product of the action of B. aceti on dextrose is
gluconic acid. If we consider dextrose, as it is generally taken, to
have the constitution of an aldehyde of the hexhydric alcohol,
mannitol, the reaction would probably be represented thus :—

Dextrose. Gluconie acid.

( CH,OH (CH,OH

Gluconic acid being the sole product of the oxidising action of
B. aceti on dextrose, considerably strengthens the idea of the
constitution of dextrose being partly aldehydic, as represented in the
above formula.

B. aceti on Sucrose.

My next experiments were made to ascertain what effect B. aceti
might have on solutions of cane-sugar or sucrose. Having failed to
make the ferment grow in solutions of cane-sugar and Pasteur’s
mineral medium, I prepared and sterilised a solution of yeast water
containing 4 per cent. of cane-sugar, and inoculated it with B. aceti
as usual. In a few days, the ferment was growing strongly. After
six weeks, the flask was opened and the solution examined. It was
found to be perfectly free from acid, and did not reduce Fehling’s
solution.

Another experiment was made in which the polarising power of the
solution was noted previous to inoculation. After B. aceti had grown
freely in the solution for three weeks, the flask was opened. The

F3 '
a |
_.
| id
ec. |
ce |
a '
ey
-s FF
Ba’ ff
a ;
Ee” 7

ve os &
.

BROWN: THE CHEMICAL ACTION OF

contents were carefully made up to the original volume and filtered.
On examining the solution with the polariscope, the original angle
was obtained. A blank experiment made with the same yeast water
and inoculated at the same time, showed that both before and after
the growth of B. aceti there was no action on polarised light. The
cane-sugar solution in my second experiment had a slight alkaline
reaction, and did not reduce Fehling’s solution ; but on treating it
with dilute sulphuric acid and warming, it reduced freely.

The above experiments show that B. aceti is unable to break up or
change the molecule of cane-sugar.

Now if the constitution of cane-sugar is really aldehydic, as it is
usually represented in the formula 04 GlH'(OH COR it appears
strange that B. aceti is unable to attack the aldehyde part of the
molecule in this case, in the same way that it acts upon dextrose
when oxidising it to gluconic acid.

Action of B. aceti on Mannitol.

After my previous experiments, it appeared very desirable to study
the action of the ferment upon mannitol, the alcohol corresponding to
the aldehyde dextrose on which the ferment acts so freely.

Gorup-Besanez (Ann. Pharm., 118, 273), on oxidising mannitol by
means of platinum black, found that mannitic acid, mannitose (a
fermentable sugar), and an unfermentable gummy substance were the
chief products. Mannitic acid he analysed and described, but he was
unable to separate mannitose from the gummy product. He describes
mannitose as not producing the least deflection of the plane of polari-
sation, but as reducing Fehling’s solution freely.

Berthelot (Ann. Chim. Phys., 50, 369) states that by fermenting a
10 per cent. solution of mannitol in water, in contact with certain
animal membranes, a “‘ glucose” is sometimes formed. The results
are very irregular, the “glucose” found varying from a mere trace
to a tenth of the amount of mannitol used, the highest result obtained.
The “glucose” could not be obtained pure, but Berthelot describes
it as uncrystallisable, very soluble in alcohol, fermentable, as reducing
Fehling’s solution, and probably levorotatory. The animal membrane
used in the experiment is thought to be the cause of conversion of
mannitol into sugar (?).

A. Fitz (Ber., 9, 1352) states that during the schizomycetic fer-
mentation of mannitol, normal butyl and ethyl alcohols, and butyr.c
and lactic acids are formed.

Hecht and Iwig (Ber., 14, 1760), by the oxidation of mannitol by
alkaline potassium permanganate, obtained formic, oxalic, and a small

PURE CULTIVATIONS OF BACTERIUM ACETI. 183

quantity of tartaric acid, and also a sugar which reduces Fehling’s
solution, and is probably mannitose.

F. Dafert (Ber., 17, 227) states that the products of the oxidation
of mannitol vary according to the agevt used, and also the time and
temperature of oxidation. Carbonic anhydride, water, formic, man-
nitic, saccharic, tartaric, and probably glycollic acids have been
observed; also a mannitose-like substance, and a sugar, mannitose,
are invariably formed. Mannitose reduces Fehling’s solution, and is
optically inactive.

My first experiment on the oxidising action of B. aceti upon man-
nitol was made with a 2 per cent. solution of the latter body (Pasteur’s
mineral medium and a little gelatin being added as food for the
ferment). After inoculation with B. aceti, the ferment grew freely.
In six weeks’ time, the fermented solution was examined. No acid had
been formed, but the solution had acquired a very sweet taste and
reduced cupric oxide freely, a result which pointed to some sugar
having been formed from the mannitol during fermentation.

In order to separate this substance, the solution was evaporated on a
water-bath and the residue exhausted with boiling alcohol, sp. gr. 830.
The part undissolved by the alcohol consisted chiefly of gelatin, and
possessed no cupric oxide reducing power. The hot alcoholic solution
(which reduced Fehling’s solution freely), on cooling, deposited some
crystals of unaltered mannitol; these were filtered off, and the clear
solution evaporated. A brown, very sweet syrup was left, which was
treated with hot absolute alcohol, in which it dissolved completely,
but on cooling a trace of mannitol and a little gummy matter were
deposited. The clear solution was separated from this deposit and
evaporated. After completely expelling all the alcohol from the
syrupy residue, it was dissolved in water and decolorised by animal
charcoal. A colourless, very sweet solution was thus obtained. On
examining with the polariscope, the matter in solution was found to
possess an [a];= — 72:0, whilst the cupric oxide reducing power was
found to equal «75°23. On adding a little yeast to the solution, it
fermented slowly, and in a few days all its optical activity had dis-
appeared, but the unfermented residue still reduced Fehling’s solution
slightly.

It is evident from the above experiment that during the growth of
B. aceti in a solution of mannitol, a fermentable sugar is formed,
possessing a high levorotatory power together with a high cupric
oxide reducing power. Levulose is the only sugar at present known
that possesses these properties.

In order to study this interesting reaction more closely, a solution
was prepared containing 25 grams of mannitol in 1 litre of yeast
water. After sterilisation and inoculation with B. aceti as usual, the

184 BROWN: THE CHEMICAL ACTION OF

fermentation appeared to go on more briskly than in my first experi-
ment. At the end of five weeks, when the flask was opened for ex-
amination, the solution contained no free acid. After evaporating the
solution to a syrup, at a temperature of 70°, it was treated with
alcohol as in my first experiment. No mannitol, however, was
found, showing that the whole 25 grams had been decomposed during
the fermentation. The brown syrup, purified as in my first experi-
ment by absolute alcohol, was further treated with cold methylic
alcohol, sp. gr. 816. In this, however, it was completely soluble. After
evaporating this solution to expel the alcohol, the syrup remaining
was dissolved in water and decolorised by animal charcoal as before.
The optical activity and the cupric oxide reducing power were then
determined with the following result :—

[a Joe = — 86-07
K336 — 86°11

In this experiment, therefore, the proportion of the rotatory power to
the cupric oxide reducing power is much the same as in my first ex-
periment, but the actual amount of levorotary sugar is much greater.

In order to see if the sugar (presumably levulose) which we have
here could be further purified, a solution of the syrup was made in
alcohol 820 sp. gr., and then excess of ether was added so as to frac-
tionally precipitate the substances in solution. Three fractions were
thus obtajned, and the cupric oxide reducing power and optical activity
were determined in each, with the following results :-—

lst Fraction... 4 bel 87-3

Qnd Fraction .. 4 Lele me anos

386
3rd Fraction... 5 Lede = ae (?).

These experiments point very strongly to the conclusion that the
sugar that has been formed from mannitol is levulose, the last fraction
especially having a rotatory power closely approximating to the sup-
posed rotatory power of pure levulose, viz., [a]; — 106.

I next endeavoured to ascertain the true rotatory power and cupric
oxide reducing power of the levulose by means of a fermentation ex-
periment. The cupric oxide reducing power and rotatory power of a
solution of the sugar of known specific gravity were carefully deter-
mined and the solution was then fermented, precautions being taken
to prevent loss of alcohol. After fermentation, the alcohol was distilled,
and the proportion of sugar decomposed was calculated from the
alcohol found (Pasteur’s figure 48°5 grams alcohol = 100 grams

PURE CULTIVATIONS OF BACTERIUM ACETI. 185

C.H,.0, sugar being used in this calculation). After determining the
rotatory power and reducing power of the residue from the fermenta-
tion, the [a | jss and Ks. of the fermented sugar were calculated from
the figures thus obtained, with the following result :—

[a]jc = — 99°8 at 15°5° C.
kos «= = 9426

Owing to at least one doubtful factor having to be used in the
above calculation (viz., the proportion of alcohol equal to sugar fer-
mented, which is unknown for levulose), the figures just given
cannot be considered to represent with great accuracy the [a]; and « of
the levulose from mannitol. Although these figures differ slightly,
therefore, from those usually assigned to the levulose from invert
sugar (viz., [a]; — 106 and « = 100), there is little or no doubt that
the two levuloses are identical.

The following properties of the levulose from mannitol, which
agree exactly with the properties of levulose both from inulin and
invert sugar, most strongly confirm this idea.

At ordinary temperatures, the sugar is a colourless, very sweet
syrup, quickly turning brown when heated at 100°. It is slightly
soluble in absolute alcohol in the cold, but on heating dissolves freely.
On cooling the solution so formed, the sugar again falls out as a
syrup; on further cooling below 0°, the syrup solidifies and becomes
opaque, bus so far I have not been able to obtain crystals of either
this sugar or the sugar from inulin, after the manner described by
Jungfleisch and Lefranc (Compt. rend., 93, 547). On treating an
aqueous solution of the levulose from mannitol with calcic hydrate,
according to Dubrunfaut’s process, a pasty mass of microscopic
crystals of a lime salt is formed resembling exactly that obtained
with the levulose of invert sugar.*

After completely fermenting aqueous solutions of the levulose from
mannitol, there is always a small unfermented residue left which has
no perceptible action on polarised light, but reduces cupric oxide.
An estimation of its reducing power after allowing for products of
fermentation left in solution, gave «ss. = 33°4. The proportion of
this compound to the levulose in the syrup from a mannitol fermenta-
tion varies, but is generally about 15 per cent. How far it has been
formed during the original fermentation, or how far from the decom-
position of levulose during evaporation of solutions, &c., is doubtful.

The experiments I have described show that mannitol is completely
oxidised by B. aceti, and that the main product formed is levulose.

* Since writing the above, I have converted levulose from mannitol into mannitol
again by means of sodium amalgam. Leevulose both from inulin and from invert
sugar is converted in a similar manner into mannitol.

VOL. XLIX.

186 BROWN: PURE CULTIVATIONS OF BACTERIUM ACETI.

In its action apon mannitol, therefore, B. aceti behaves differently
from all other oxidising agents which have been described (see above).
Previous investigators have always found a small quantity of a sugar
amongst the products of oxidation of mannitol, but this sugar is an
optically inactive one, mannitose.

If the constitution of levulose be represented by the formula
which Kiliani (Ber., 18, 3066) has recently shown to be probably the
correct one, the following will perhaps represent the action of B. aceti
upon mannitol :—

(CH,;OH (CH,;OH

| CH.OH | CHOH
CH-OH _ JO

+ OOH + 0 | Con + OH,.
CH-OH CH-OH

| CH,-OH | CH,-OH

Mannitol. Leevulose.

In my experiments, one fact is very noticeable, viz., that no acid is
formed during the decomposition of mannitol. Fermentations left for
a month after the complete disappearance of all the mannitol showed
no trace of acid. From this it was evident that the levulose formed
could not be further oxidised by the ferment to gluconic acid in
the same way as dextrose. Experiments made by growing B. aceti
in solutions containing levulose prepared from inulin showed that
this sugar also was not oxidised by the ferment. This in itself is
evidence that there must be a very considerable difference between the
molecular constitution of lzvulose and dextrose, and appears to
strengthen the theory that the one is a ketonic compound, and the
other aldehydic. Thus,

(CH,OH)

CH:OH |
a } co CO | Bot oxidised by
~"* \ CH-OH { B. aceti.
| CH-OH |
l CH OH _)

(CH,OH )
| CH-OH fee
: CH-OH { Oxidised by
Dextrose 4 CH-OH rR. anite 7
CH-OH |
COH ) (COOH J

As we know now that by means of B. aceti we can convert mannitol
into levulose, it follows that dextrose can be converted into levulose
through this reaction, by first transforming it into mannitol by means

eo Vw VP Aa ete 8. Re OS SSlU

STALLARD: THE MONOBROMOPHTHALIC ACIDS. - 187

of sodium amalgam, always supposing that the mannitol so formed is
identical with that from manna.

I think the experiments just described will be of interest to biolo-
gists, as well as chemists, as they help to show that the vital functions
of certain organised ferments are most intimately connected with the
molecular constitution of bodies upon which they act.

My best thanks are due to Mr. Horace T. Brown, and to Dr. G. H.
Morris for help given me in various ways whilst pursuing this inves-
tigation.

XX.—The Monobromophthalie Acids.
By G. Sratiarp, M.A.

Two monobromophthalic acids have been hitherto described, the
first obtained by Faust (Annalen, 160, 62) by the action of bromine
and water on phthalic acid at 180—200°, the second by Guareschi,
Meldola, and Smith, by the oxidation of certain bromo-derivatives of
naphthalene. The first acid and its anhydride are said to melt at
138—140° and 60—65° respectively, the second at 174—176° and
133—135° respectively.

The two theoretically possible monobromophthalic acids are repre-
sented by the formule —

Br
/™\ COOH Br “ \ COOH
and
COOH COOH
IL. Il.

Faust’s acid is supposed to correspond to formula I on the strength
of an observation of H. v. Pechmann’s (Ber., 12, 2126) that this acid
can by a series of processes be converted into erythroxyanthra-
quinone. Both Guareschi and Meldola accordingly assume that the
acid melting at 174—176°* is that represented by the second symbol.
The force of these assumptions is, however, much diminished by the
fact that the action of bromine and water on phthalic acid, according
to v. Pechmann, gives rise to both the theoretically possible mono-

* Carnelley (Chem. Soc. Trans., 1885, 591) also describes a bromophthalic acid
obtained by the oxidation of a bromoditolyl, and which, though it melts at 203°,
he regards as identical with Guareschi and Meldola’s acid.

188 STALLARD: THE MONOBROMOPHTHALIC ACIDS.

bromophthalic acids, and that he confessedly obtained his erythroxy.
anthraquinone from a mixture of these acids, not having succeeded in
separating them from one another.

Now Guareschi (Annalen, 222, 292 and 298) obtained the mono-
bromophthalic acid melting at 174—176° by oxidation, first, of a bromo-
nitronaphthalene, C,yH,Br,NO,., obtained by the action of bromine on
nitronaphthalene in molecular proportions, and, secondly, by oxida-
tion of the corresponding bromonaphthylamine, C,H,Br-NH:. He
therefore assumes that the constitution of these compounds is
correctly represented by the symbols

*Br

in other words that the bromine-atom enters the second “ ring” ina
B-position.

I have recently prepared these compounds (each of which is
undoubtedly a single substance, and not a mixture of isomerides)
according to Guareschi’s directions, and found that they possess the
melting points and other properties assigned to them by that author;
and with the view of throwing light upon their constitution, and that
of the bromophthalic acid related to them, have made the following
experiments.

The bromonaphthylamine was converted in the ordinary way into the
diazo-sulphate, and this was decomposed with absolute alcohol. The
alcohol was then distilled off, and the sulphonic acid obtained sub-
jected to steam distillation, the temperature of the liquid being kept
at 135—140°. The slightly yellow oil thus obtained was unaltered
in appearance by a second steam distillation, had a sp. gr. of 15, did
not solidify at — 15°, and proved to be «a-bromonaphthalene.
Nothing further was obtained by even a very considerable rise in the
temperature at which the steam disiillation was effected.

To another portion of the diazo-sulphate excess of bromine-water
was added, and the flocculent orange precipitate of diazo-perbromide
was warmed with glacial acetic acid till evolution of nitrogen ceased.
From the acetic acid solution, a small quantity of water precipitated
a substance which, after solution in slightly diluted alcohol with the
addition of animal charcoal, gave fine brilliant needles of a dibromo-
naphthalene melting at 130—130°5°. Water was then added to the

* I omit, in this and similar cases, the alternative formula, since chemists have
at present no means of ascertaining by which of two such symbols a given body is
more correctly represented.

STALLARD: THE MONOBROMOPHTHALIC AOIDS. 189

acetic acid mother-liquor as long as a precipitate was formed; this
fraction proved to be identical in composition and melting point with
the first,

This dibromonaphthalene crystallises also in needles from glacial
acetic acid, and appears to be identical, in spite of the difference in
crystalline form, with the so-called y-dibromonaphthalene obtained by
Jolin by distilling a-dinitronaphthalene with phosphorus penta-
bromide, and found by Guareschi and Magatti among the products
of the action of two molecular proportions of bromine on naphthalene ;
it is also probably identical with the compound melting at 126—127°,
obtained in the form of microscopic needles by Dairmstadter and
Wichelhaus (Annalen, 152, 303) by the action of 1 mol. of bromine
on a-C,,H,SO;H.,

It appears, therefore, that this dibromonaphthalene contains, as has
long been supposed to be the case, its two bromine-atoms in a-posi-
tions, and one in each nucleus, since the isomeric /A-dibromo-
naphthalene (m. p, 81°), which also has its bromine-atoms in a-posi-
tions, contains them in the same nucleus. If this be so, Guareschi’s
bromonitronaphthalene must have a similar constitution, which may

NO,

A\/\
be expressed by the formula [ | |. and the bromophthalic acid

/
Br .

melting at 174—176° must have its bromine-atom in the position
contiguous to the two carboxyl-groups. This view is confirmed by a
recent paper of Kriiger’s (Ber., 18, 1755), in which he shows that the
anhydride of the monochlorophthalic acid obtained by the oxidation
of the 1: 2: 3 monochlorortho-xylene has a much higher melting point
than that from the isomeric 1 : 2: 4 derivative of the same hydro-
carbon. On the other hand, it must be admitted that Meldola’s
experiments (Trans., 1885, 511—515) favour a view exactly opposite
to that here suggested as to the constitution of the monobro-
mophthalic acids. This author, by bromination of the two bromo-
naphthylamines of the formulze

\\ NH, ry Br
and | | ,
FT
NH;

Br
} a

has obtained two dibromonaphthylamines, giving on hydrolysis the

same dibromonaphthalene (m. p. 75°), which is, therefore, an a-A-

modification. The bromine-atoms, moreover, must be in different
0 2

190 MCGOWAN: SOME DERIVATIVES OF THIOCARBAMIDE,

nuclei, since the formula OO is already assigned with good
Br
*

reason to the compound described by the same author in the Trans.,
1883, 5. The dibromonaphthylamines in question are, therefore,
represented by the formule

Br AN NH, A\/\
| | and L
ne 4 Br ‘E,

I. II.

respectively.

The first of these gives on oxidation (according to Meldola)
the bromophthalic acid melting at 174—176°; from the second he
failed to obtain a bromophthalic acid, and unless we suppose that
during the oxidation process an atomic migration occurs, it neces-
sarily follows that the usual assumption as to the constitution of the
bromophthalic acids is correct.

It appears, however, that in the oxidation of naphthalene-deriva-
tives there may be, and in some cases probably are produced, owing to
secondary actions, substituted phthalic acids other than those which
might naturally be expected. If this be so, it is necessary to receive
with caution a formula for a naphthalene-derivative based upon the
results of an oxidation experiment, and a far safer method than that
commonly used in oxidising haloid substitution derivatives of this
hydrocarbon would appear to be that adopted by Claus (Ber., 15,
320) in the case of e-dichloronaphthalene, viz., to perform the experi-
ment in presence of silver nitrate with the view of preventing

secondary action.

XXI.—Some Derivatives of Thiocarbamide.

By Grorce McGowan, Ph.D., F.R.S.E, Demonstrator in Chemistry,
University College, Bangor.

In continuation of my work on the derivatives of methylsulphonic
acid (J. pr. Chem. [2], 30, 280), I studied at the suggestion of Herr
O. Loew (in a private letter written after his having read the paper
just mentioned) the action of thiocarbamide on trichloromethyl-

MCGOWAN: SOME DERIVATIVES OF THIOCARBAMIDE. 191

sulphonic chloride, CCl, : SO,Cl, and from this I have been led on to
investigate certain thiocarbamide compounds.

If the above-mentioned sulphonic chloride is added to an aqueous
solution of thiocarbamide, no reaction takes place in the cold until
after long standing, but, on warming, the chloride gradually dissolves,
and sulphur is deposited. If, however, a cold or nearly cold solution
of thiocarbamide in 80 to 90 per cent. alcohol is used, there is con-
siderable development of heat on adding the chloride, and a mass of
needles separates; the quantity of the latter being much increased by
the subsequent addition of ether.

This compound was found to be identical with that obtained by
Claus (Annalen, 179, 139) by the action of chlorine on an alcoholic,
or (with careful cooling) aqueous solution of thiocarbamide, viz.,
dithiocarbamide dichloride, (CSN,H,).Cl. In preparing it as above
from trichloromethylsulphonic chloride, it is best to take quantities
of the pure substances representing two molecular proportions of the
sulphonic chloride and four of the thiocarbamide. The latter being
dissolved in a small quantity of 90 per cent. alcohol, the chloride is
gradually added, care being taken to avoid much rise of temperature.
When the reaction is complete, an equal volume of ether is added,
and, after the whole has stood for a short time, the dithiocarbamide
dichloride may be collected and washed, first with a mixture of
absolute alcohol and ether, and then with ether, and finally dried over
sulphuric acid in a vacuum.

The compound thus prepared yielded the following results on
analysis :—

(a.) 0°1590 gram gave 33°3 cc. N at 16° and 770 mm. mercury
pressure = 24°75 per cent. N. .

(b.) 0°2841 gram ignited with copper oxide and lead chromate
gave 0°1076 gram CO, and 0°0995 gram H,O = 10°33 per
cent. C and 3°89 per cent. H.

(c.) 0°3280 gram gave 0°1238 gram CO, and 0°1166 gram H.0 =
10°29 per cent. C and 3°95 per cent. H.

(d.) 0°1857 gram ignited with pure lime gave 0°2371 gram AgCl
= 31°59 per cent. Cl.

Calculated for
(CSN,H,)Cl,.
3°95
10°33 10°29
24°75 —
31°59 —

A compound CSN,H;Cl would require 2°72 per cent, H,

192 MCGOWAN: SOME DERIVATIVES OF THIOCARBAMIDE,

The filtrate from the above dithiocarbamide dichloride contains
another product crystallising in stellate groups of the finest needles,
and having probably the composition (CSN,H,),(CC1;,SO,).. This
will be referred to in a future paper.

Under similar conditions, methylsulphonic and dichloromethy]l.
sulphonic chlorides, CH;'SO,Cl and CHCl,SO,Cl, yield the same
product, as does also phosphorus oxychloride, POCI;;* but I failed
to obtain it with acetyl chloride.

As already described by Claus, the above dithiocarbamide dichloride
crystallises in rather small colourless needles, readily soluble in water,
less so in alcohol, and insoluble in ether.

If a small portion is dissolved in 90 per cent. alcohol on a micro-
scopic slide, the compound separates out again (after the slow evapo-
ration of the alcohol) in highly characteristic plates (monoclinic ?).
It also crystallises in the same way from its cold aqueous solution,
but in this case the plates are very much elongated and mixed with
long needles.

In like manner, the dichloride, dissolved in water on a microscopic
slide, gave—

(a.) With dilute nitric acid, a crystalline precipitate, the crystals
being like jagged daggers or long sword-like leaves (see
below).

(b.) With platinum tetrachloride, large thin rectangular crystals.
I have not investigated this compound.

The very interesting reaction which this dichloride (or the corre-
sponding dibromide) gives, on heating its aqueous or alcoholic
solution, or on the addition of an alkali, has already been explained
by Claus, according to the following equation :—

(CSN,H,).Cl, = 2HCl + Ss + CSN.H, + CNNH,,

Thiocarbamide. Cyanamide.

or a derivative of cyanamide, e.g., dicyandiamidin. He was unable
to obtain the cyanamide free from thiocarbamide for analysis; how-
ever, exactly the same reaction has quite recently been observed
by Verneuil (Compt. rend., 100, 1296), for the corresponding
selenium compound, (CSeN,H,).Cl,, which sufficiently corroborates
his view.

Dithiocarbamide dichloride cannot be recrystallised from water or
alcohol, as the solutions invariably deposit sulphur. An aqueous
solution strongly acidified with hydrochloric acid may, however, even
be heated with impunity, and in the cold the dichloride separates out

* This phosphorus oxychloride was not quite pure.

MCGOWAN: SOME DERIVATIVES OF THIOCARBAMIDE. 193

in large thin plates (similar to those on the microscopic slide
mentioned above) ; but I have never, even in this way, been able to
obtain it quite free from deposited sulphur. Hence the necessity of
preparing it pure in the first instance. It is stable at ordinary
temperatures.

Action of Dilute Nitrie Acid on Dithiocarbamide Dichloride.

If, as already indicated, dilute nitric acid is added to a moderately
concentrated aqueous solution of the dichloride, most characteristic
colourless crystals separate at once, while the liquid is found to
contain free hydrochloric acid. These crystals are almost insoluble in
dilute nitric acid, more soluble in water (but still only sparingly),
practically insoluble in cold absolute alcohol, and insoluble in ether.
The compound is thus easily obtained pure by washing successively
with water, alcohol, and ether, and drying over sulphuric acid in a
vacuum.

These crystals were also observed to be formed in the mother-
liquor from the preparation of thiocarbamide nitrate : CSN,H,,HNO,;
(formed by adding strong nitric acid to a concentrated aqueous solu-
tion of thiocarbamide), red fumes being at the same time evolved.

The pure compound gave the following results on analysis :—

(a.) 0°3783 gram gave 0°1228 gram CO, and 0°1064 gram H,O =
8°85 per cent. C and 3°12 per cent. H.

(b.) 04097 gram gave 0°1232 gram CO, and 0°1138 gram H,O =
8°20 per cent. C and 3°09 per cent. H.

(c.) 0°1496 gram gave 37°6 c.c. N at 12° and 771 mm. = 30°27
per cent. N.

(d.) 0°1510 gram gave 38°2 c.c. N at 11° and 766 mm. = 30°41
per cent. N,

Calculated for Found.
(CSN2H,)2(NO3)2. pee a,
88: 8°20
31: 3°09
30°41

The nitrogen estimation (d) was made from a sample 10 days old.
The two estimations of carbon and hydrogen were made from the
same preparation.

Thiocarbamide nitrate, CSN,H,,HNOs, requires 3°61 p. c. H.
A compound, CSN,H;NO;, would require 2°19 p. c. H.
CSN.H;NO,, + 2°48 -

”

194 MCGOWAN: SOME DERIVATIVES OF THIOCARBAMIDE,

This somewhat peculiar compound is therefore dithiocarbamide
dinitrate, (CSN,H,)(NOs)>.

A further confirmation of its composition is given by the filtrate
(from the dichloride and dilute nitric acid) being found to contain
much free hydrochloric acid, but not a trace of chlorine, thus :—

(CSN;H,).Ch + 2HNO, = 2HCl + (CSN,H,).(NO;)>.

That this filtrate contains extremely little dinitrate in solution is
easily proved by the addition of a solution of silver nitrate. First
silver chloride is thrown down, and then only a very small quantity
of a silver thiocarbamide compound.

The above dinitrate is comparatively stable, and therefore differs in
this respect from thiocarbamide nitrate, as described by Reynolds
(Annalen, 150, 231). Possibly the extreme instability of Reynolds’
compound may have been due to the presence of a small quantity of
free nitric acid.

Although a good many samples of the dinitrate were prepared,
and kept for months in tubes over calcium chloride, only one
exploded, and that one almost immediately after preparation (I
noticed that in this case the crystals were very small, i.e. the
compound was very finely divided). Jt appears to withstand a
temperature of 60° for some hours with little or no decomposition,
but decomposes gradually on being heated more strongly. When
kept in a dry tube, decomposition goes on very slowly, hydrocyanic
acid being evidently evolved and sulpbur deposited. It separates
from the dilute mother-liquor of its preparation in beautiful long
columnar prisms. Like the dichloride, it at once deposits sulphur
when its solution is heated or on the addition of an alkali to it.

A solution of potassium nitrate likewise precipitates at once a cold
aqueous solution of dithiocarbamide dichloride. The crystals in this
case are somewhat different in form; they are long well-developed
prisms (possibly monoclinic), among them being many well-crystallised
individuals. From dilute nitric acid the crystals are more
agglomerated.

Action of the Chlorides of Iodine on Thiocarbamide.

If iodine monochloride, ICI, or trichloride, ICl;, is added to a cold
solution of thiocarbamide in alcohol, dithiocarbamide dichloride 1s
formed, and iodine separates, even when the thiocarbamide is present

in excess :—

2CSN.H, + 2IC] = (CSN,H,).Ch, + kh.

MCGOWAN: SOME DERIVATIVES OF THIOCARBAMIDE. 195

With a concentrated solution, or with powdered thiocarbamide
moistened with alcohol, the action is very violent.

In aqueous solution, the iodine seems to separate more slowly,
indeed in an aqueous solution containing an excess of thiocarbamide,
a certain amount of iodine compound seems to remain undecomposed
until the solation is heated (comp. Verneuil on the seleniothio-
carbamide compounds, Compt. rend., 100, 1296).

Action of Potassium Iodide on Dithiocarbamide Dichloride.—If
equivalent quantities of these (finely powdered) be shaken up
together in absolute alcohol, they react as shown in the equation :—

(CSN,H,).Cl, + KI a 2KCl + 2CSN.H, + I.

Here again the separation of iodine appears to go on more slowly if
an aqueous solution is used.

Dithiocarbamide Di-iodide, (CSN2H,)2I..

This can, however, be prepared by rubbing together in a mortar
pure thiocarbamide and iodine in the proportions of their molecular
weights, just sufficient alcohol being added to make a pasty mass.
In this way the iodine is taken up, and a compound crystallising in
long prisms formed. The reaction succeeds best with small
quantities; for instance, 0°47 gram iodine and 0°28 gram thio-
carbamide. The product is treated several times with cold benzene
to remove traces of free iodine, and the nearly colourless residue is
then dried between folds of filter-paper.

03528 gram gave 39°8 c.c. N at 12° and 768 mm. = 13°54 per
cent. N.

(CSN,H,)I, requires 13°79 per cent.

When prepared, this di-iodide is quite stable and can be kept for
any length of time in a dry tube. On being warmed it melts—at a
comparatively low temperature—with liberation of iodine. It
crystallises, as already mentioned, in long colourless prisms. There
is some difficulty in getting good crystals for the microscope.

Benzene does not dissolve it, but water, alcohol, and ether do, with
immediate separation of iodine, which accounts for its not being formed
in solution. This decomposition of the di-iodide occurs even if it be
dropped into absolute alcohol which has been cooled by a mixture of
ice and salt, or on adding a drop of absolute alcohol to the di-iodide
covered by a layer of benzene ; chlorine and concentrated hydrochloric
acid added to it likewise liberate iodine at once.

On heating the aqueous solution, or on addition of caustic alkali,

196 MCGOWAN: SOME DERIVATIVES OF THIOCAKBAMIDE.

sulphur separates. This separation of sulphur is therefore charac-
teristic of the di- (chlor-, brom-, nitro-, §c.) derivatives of thiocarbamide.

Dilute hydrochloric acid does not cause separation of iodine in
the cold (indeed it seems insoluble in this and also in dilute nitric
acid) ; on heating the mixture, however, iodine is liberated freely, but
no precipitation of sulphur occurs. This shows that it is decomposed,
under those conditions, more readily into iodine and thiocarbamide
than into hydriodic acid, sulphur, thiocarbamide, and cyanamide
(compare with the decomposition products of the dichloride or dibro-
mide). Although the di-iodide appears to be insoluble in diluie
nitric acid, still, if a drop of alcohol be added to a little of it ona
microscopic slide, and then at once a drop of dilute nitric acid, the
dinitrate is precipitated.

With regard to the action of cyanides on thiocarbamide and its
dichloride, I need merely say at present that it appears to be
impossible to prepare dithiocarbamide dicyanide, as this substance
splits up immediately after formation. I hope, however, to go more
fully into this point in a future paper.

University College of North Wales,
Bangor, December, 1885.

XXII.—Bacteriological Research from a Biologist’s Point of View.

By E. Kuziw, M.D., F.R.S., Lecturer on General Anatomy and
Physiology in the Medical School of St. Bartholomew’s Hospital,
London.

THE memorable and classical researches of Pasteur on the relation
of micro-organisms to various fermentative processes and to putrefac-
tion have given an active impetus to investigations on the chemical
activity of micro-organisms, and during the last 10 or 15 yearsa
great number of facts have been brought to light, showing how
enormous and important the réle is that bacteria and yeasts play in
the economy of nature (comp. Professor Frankland’s paper, Trans.,
1885, 159). A great deal of this knowledge is real, and has been
proved to be so by numerous accurate experiments. Thus it has
been shown beyond any doubt that alcoholic fermentation is caused
by the multiplication and activity of the yeast, known as Saccha-
romyces, that acetic acid fermentation is produced by Mycoderma
aceti, that lactic acid fermentation is produced by Bacterium lactis,
and it is also known that putrefaction is caused by a variety of
species of bacilli. A good deal, however, of knowledge obtained with
regard to chemical processes produced by micro-organisms, although
put forward as perfect, om close examination is found to be in a very
imperfect state.

I will illustrate this by a few examples, which I could multiply.
There can be no doubt that the complex process spoken of as putre-
faction of proteids is due to bacteria, but when we come to inquire
which part of the process is due to which bacteria, we receive no
definite answer. Putrefaction is a process by which proteids undergo
changes, beginning with the formation of peptones, leading then to
the formation of leucine and tyrosine, of indole, skatole, phenol, and
a variety of substances belonging to the aromatic series, and further
of certain alkaloids known as ptomaines, ammonia and its salts, and
nitrates, with the simultaneous development of sulphuretted hydrogen.
Now no one has yet shown whether this whole series of changes is
due to one kind of organism, or whether one kind of organism
commences the process, and it is then carried on a step further by
another kind of organism. We have certain well-established experi-
ments by which it is shown that a certain kind of organism produces
the change of sugar into alcohol, that is, the Saccharomyces cerevisie.
After this has finished converting sugar into alcohol, another orga-
nism steps in and changes the alcohol into acetic acid. It is probable

VOL. XLIX. P

198 KLEIN: BACTERIOLOGICAL RESEARCH

that a similarly complex process occurs in putrefaction, that is to
say, it is probable that one set of organisms only brings the proteids
down to a certain stage, preparing as it were the ground for another
set which then begin their activity.

Another illustration which I can give of the imperfect state of our
knowledge is this. The ammoniacal fermentation of urine,
consisting as you all know in the conversion of urea into ammonium
carbonate, is according to the account given by Pasteur and Cohn,
due to a micrococcus called the Micrococcus uree. Professor Frank-
land told you in his address (loc. cit., p. 178), that he found in
ammoniacal urine a kind of vibrio which he considers to be the
ferment of the ammoniacal change. Now I have myself examined a
great many samples of urine that have undergone ammoniacal fermen-
tation, but I have not found this vibrio. Had Professor Frankland
isolated this vibrio, and, having isolated it by pure cultivations, had he
then produced with it the ammoniacal fermentation in sterile urine,
we should have been able to say that his so-called Bacillus uree is
also an active agent in the conversion of urea into ammonium carbo-
nate ; but not having done so, and merely finding it present in ammo-
niacal urine, the assertion that it is the cause of the ammoniacal
fermentation is open to question. To illustrate, at the same time, how
an investigation of this nature ought to be carried out in order
to command absolute value, I will mention to you experiments
which were conducted last year by Leube and Graser (Virchow’s
Archiv, 100, 3). These gentlemen set themselves the task to
determine which organisms produce the ammoniacal fermentation of
urine. They proceeded in this way. They took ammoniacal urine,
that is to say, urine after it had undergone ammoniacal fermentation,
and isolated by careful experiments, in a way which I shall describe
more minutely later on, the various bacteria present in such urine.
Amongst the great variety of bacteria present, four kinds proved
themselves capable of producing the ammoniacal changes of urea ; first
and foremost, a sort of straight bacillus, thick and rounded at its ends;
secondly, a micrococcus identical with the known Micrococcus wree;
the third, but of less pronounced activity, is a small short bacillus, oval
in shape; whilst the fourth, still less active, is a very short bacillus,
with truncated ends. These four kinds of organisms have been isolated,
carried in pure cultivations through many successive generations,
their morphological characters well ascertained, so that they might
be recognised at once whether present or absent in a given cultiva-
tion ; and with each species separately, sterile urine or sterile fluids
containing urea were inoculated, and the chemical action of the
several species studied.

To illustrate, on the other hand, an inconclusive experiment, I will

FROM’ ‘A BIOLOGIST’S POINT OF: VIEW. 199;

mention the following examples: We are told by Professor Frank-
land that, according to Bell, Mucor racemosus is capable of producing:
alcohol out of sugar, that is to say, that it is also an alcoholic ferment.:
Now as far as I am able to say, this Mucor racemosus has not been
isolated by exact methods; all that has been ascertained is that
Mucor racemosus has been sometimes found to be present where
alcoholic fermentation has been going on. But if you come to:
inquire more carefully into the matter, you will find that besides:
Mucor racemosus, Saccharomyces has also been present, and one would
say with little hesitation that it was the latter organism which pro-:
duced the alcoholic fermentation, whilst the mucor was merely present
accidentally. A few years ago, Dr. Kern (Biologisches Centralblatt, Il),
described a peculiar bacillus, which owing to its containing a spore at
each end he named Dispora, and because it was found in fermenting
milk, used in the Caucasus as a drink under the name of “ kephir”
or “ hippé,” it was termed Dispora Caucasica, Now of this Dispora
Caucasica Kern maintained that it is the active ferment in this alco-
holic fermentation. Why did he maintain that? Because he found
it always present in fluids undergoing this fermentation. Later on
he was able to isolate it by pure cultivations, and then he became
convinced that this bacillus had nothing to do with the alcoholic
fermentation, but that the real agent in this process is a Saccharo-
myces; consequently this Dispora Caucasica is merely an accidental
concomitant of this particular process. Similar criticisms might be
applied to the various organisms which have been mentioned as the
causes of the butyric fermentation by Fitz and Bell. As far as
at present known, Van Tieghem and Prazmowski’s Clostridium buty-
ricwm is the real ferment in this process. Many of the assertions made
as to organisms having caused certain chemical processes are there-
fore untrustworthy, because they are due to either of the following
sources of error: first, working with non-sterilised material; and,
secondly, not isolating the organisms and testing their activity. IEf of
any organism it is to be said with something like certainty that it
really is the active cause of a specific chemical process, it must be
shown, first, that when obtained pure, that is, when isolated it
possesses certain well-defined characters ; secondly, when introduced
in this pure state into a suitable material, it must set up the specilic
action. These elementary conditions of experimenting are thoroughly
appreciated and employed by most pathologists at the present day
in their investigations into the relation of disease germs, but I
am afraid they have not yet been fully understood by chemical
investigators.

I will describe to you the methods which pathologists employ in
ascertaining whether a given disease is due to a certain organism.

200 KLEIN: BACTERIOLOGICAL RESEARCH

In the first place, the pathologist does ascertain whether an organism
and which organism is present in the diseased tissues. In a variety
of ways he studies the characters of such an organism by fresh exa-
mination, by methods of staining with various dyes, and by cultures.
By this latter method, he not only studies their behaviour in various
culture media, but at the same time isolates the organism in order to
obtain pure cultivations. Having then obtained unmistakably pure
cultivations, he proceeds to inoculate with them suitable animals, and
ascertains whether this purified organism is capable of again pro-
ducing the same disease from which it had originally been derived.
And, further, in this so induced disease, the same organism must
again be detected; by morphological study, and by the method of
culture, it must become clear that in this new infection we meet with
exactly the same organism as in the original disease.

I shall describe to you then, first, the method of cultivation and
isolation, and, secondly, the method of testing the activity of the
purified organism.

Supposing we have to do with the disease known as malignant
anthrax: in this disease, the blood is the soil or locality in which the
Bacillus anthracis finds its most suitable conditions for active growth
and multiplication. The blood of an animal dead or dying of malig-
nant anthrax very often contains vast numbers of this Bacillus
anthracis. Every droplet of such blood inoculated into a suitable
animal produces with certainty malignant anthrax and death in this
animal. Its blood again teems with the Bacillus anthracis. But
notwithstanding this we do not yet conclude that this Bacillus
anthracis is the active ferment of this disease, because the Bacillus
anthracis may be only an accidental concomitant; there may be
another organised or non-organised ferment present, which is the real
virus. In order to be able to say that the Bacillus anthracis is really
the active virus, pathologists have found it necessary to isolate this
Bacillus anthracis by cultivation. Koch was the first who achieved
this cultivation. A drop of blood of an animal dead of anthrax, or a
bit of the spleen of such an animal, was introduced into some
suitable nourishing medium. He found that the Bacillus anthracis
soon undergoes rapid multiplication by division, and that after a
certain time, only hours sometimes, it formed in its interior bright
oval seeds or spores. These spores were then introduced into fresh
nourishing media, and they were seen to grow out or to germinate
into bacilli, which rapidly multiplied by division. These bacilli,
either after they had been forming spores, or before this period had
been reached, were transferred to a fresh nourishing medium, in order
to obtain a new crop of bacilli, and in this way by a number of
successive cultivations he obtained the bacilli in a pure state, that is

FROM A BIOLOGIST’S POINT OF VIEW. 201

to say, he obtained bacilli which had been far removed from their
original nidus, ¢.e., the blood. Now he found that however far
removed cultures of the bacilli were, they always have the same
action, namely, when introduced into a suitable animal, they invariably
produce malignant anthrax and death, the blood and spleen of such
animals containing vast numbers of the bacilli. In the same way,
Koch proved that the bacillus known as Bacillus tuberculosus is the
active agent of the pathological process known as tuberculosis.

Fehleisen proved by the same kind of experiments that a micro-
coccus is the real cause of erysipelas, and I could mention a number
of other instances, e.g., septicemia and swine fever, and many others,
where a definite infectious disease has been proved beyond any doubt
to be caused by a definite species of organism.

In all these experiments of pure cultivations, the material in
which the organism is to grow must at the outset be made
sterile, for if by any chance another organism, say a septic organism,
happens to be present at the outset, this will suppress the disease
organism that is introduced. The sterilisation of nourishing mate-
rials is therefore the first and I may say an essential condition for
the success of the experiment; and this—namely, the sterilisation
of the nutritive material in which the given organism is to grow, and
in which it is to be obtained in pure culture—must be the first
attempt in all experiments coming under this head. The sterilisation
of nutritive materials is always effected by heat, The same of
course applies to all vessels, instruments, &c., used in eonneetion with
such nourishing materials, the vessels that are to contain the nourish-
ing material and the instruments that are used for the putting in or
taking out the organisms. If you have to deal with a given nourish-
ing material, it is necessary after sterilisation, or after supposed steri-
lisation, to keep it under such conditions as would facilitate the
development of organisms. If the material remains unaltered, and
barren of organisms, under these conditions you are justified in
saying that this material is free from any germs. Having aseertained
that this is the case, you introduce your organism whieh you wish to
test in its chemical] activity. Of course you have to be careful to
introduce only that one organism which is to be tested, that is to say,
you have to be careful of accidental contamination with organisms
that are not wanted. There is at present a tendency to neglect acci-
dental contamination by organisms contained in the air. You see in
many pamphlets, in some books also, that it is recommended when
introducing a definite organism into fresh nutritive material, that
is to say, when inoculating nutritive material with a given organism,
to expose the nutritive material to the free aceess of air during inocu-
lation. Now there is no doubt that at certain seasons, and at some

ee ee a ee

202 KLEIN: BACTERIOLOGICAL RESEARCH

places, nutritive materials can be exposed to free access of air; of
course for short periods, say seconds, without any actual contamina-
tion with air organisms taking place. But anyone who has worked
at this kind of research at all seasons of the year, and in a variety of
localities, must have found that the danger from accidental contami-
nation with air organisms is areal one. Professor Tyndall has drawn
attention in a masterly way to these dangers, and has proved that,
however short the time may be during which suitable nourishing
materials are exposed to free air, accidental contamination with septic
organisms does take place oftener and to a larger extent than can be
desired by those who are engaged in such work. In London, free
exposure of suitable nutritive materials to air is followed in a certain
percentage of cases by contamination with the following organisms:
—foremost are the spores of moulds; then torula or saccharomyces:
then micrococcus and the spores of bacilli.

The isolation of bacteria which happen to be present in a given
fluid or solid is achieved in a variety of ways. The method which is
now employed to a great extent is that of plate cultivation, first
introduced by Koch. It is based on the fact that when a variety of
organisms are introduced into nutritive gelatin, previously liquefied
and distributed, by shaking, through this gelatin, after the gelatin is
allowed to congeal, the different organisms are kept permanently
in different places. Then the conditions under which the gelatin is
placed are such that organisms can multiply, it is found that each
one of those different organisms will give rise to a colony made
up of members of its own kind. In this way, it can soon be ascer-
tained which kind of organisms has been introduced, and even how
many different kinds have been introduced, provided that the original
number introduced is not too large, so that the different colonies
making their appearance in the gelatin remain sufficiently isolated.
This is practically the method which Dr. Angus Smith first used for
the study of bacteria in water. The better plan is that adopted by
Koch, namely, after the introduction of the bacteria mixture into the
liquefied gelatin, the latter is poured out on glass plates or similar
vessels, previously sterilised, and then kept in a closed moist chamber.
The gelatin of course is allowed to set, and is kept in the solid state,
that is to say, the temperature at which these gelatin plates are kept
must not be higher than the temperature at which the gelatin remains
solid. After a few days,a number of spots are formed on these plates,
each spot corresponding toa colony ; the different colonies, when due to
different species, show different characters (size, colour, liquefaction
or not of the gelatin, &c., &c.) already tothe unaided eye. From such
plate cultivation, pure cultures can be made by inoculating from any
of these different colonies fresh nutritive material contained in flasks

FROM A BIOLOGIST’S POINT OF VIEW. 203

or test-tubes, or any other vessel, of course the nutritive material
having been previously sterilised. This is the simplest method of
isolation, and the one at present the most employed. It has its draw-
backs, and it has its great advantages. The drawbacks are chiefly
these: first, accidental contamination with air organisms is not
excluded. You may therefore obtain a number of colonies in your
plate cultivations which have not been represented in the fluid from
which you derived your bacteria. This may not count for much
when you have to deal with the isolation of an organism the special
characters of growth of which you have previously already ascer-
tained, for you may easily see whether in the plate cultivation your
organism has made its appearance, and it makes very little difference
whether there are any others present, or how many, but it is of con-
siderable importance when you are first searching for and ascertaining
the existence of an organism the characters of which you do not yet
know.

Another disadvantage of this method is that although many organisms
are capable of growth in nutritive gelatin, such as is commonly used,
there are other organisms, and some very important organisms, at
any rate important to pathologists, which refuse to grow in gelatin.
Again, there are other organisms which, although not incapable of
growth in gelatin, will not grow in it, or far too slowly, at the tem-
perature at which gelatin remains solid; they require much higher
temperatures in order to develop into colonies that can be recognised
by the unaided eye. These disadvantages apply also to water analysis
by plate cultivations, such as have been minutely described and exten-
sively used by Dr. Percy Frankland.

Another very good method of isolation is that by “dilution.” In
this method a given number of different bacteria are distributed over
or in a large quantity of previously sterilised indifferent fluid, so that
a definite quantity of this fluid contains only one organism. With
this definite quantity of fluid nutritive material is inoculated, and
then such nutritive material becomes the seat of the growth of that
one kind of organism. This is the method which was first used by
v. Naegeli, and afterwards with great success by Lister, in his investi-
gations of Bacterium lactis.

A third method of isolation is that known under the name of
“fractional cultivation,” first used by Klebs. In this method the
fact is made use of that different organisms grow to a different
degree in different nutritive media. It is within the knowledge of
everybody who has worked with bacteria, that different bacteria
require different materials to grow in—some grow better than others
in a given nutritive medium. Now supposing we have a fluid or a
solid which contains, as the microscope proves, a variety of bacteria,

204 KLEIN: BACTERIOLOGICAL RESEARCH.

on inoculating with this mixture of bacteria one and the same kind of
simple nutritive medium contained in several test-tubes, and exposing
these to suitable conditions of growth, it will be found that after one
or two days there is present in this particular nutritive medium the
offspring, not of all the organisms that have been introduced, but
pre-eminently and predominantly, perhaps, only those of one species.
If that be the case, there is no difficulty in obtaining pure cultures
from this one species, provided we use only a trace of this growth for
the new cultivation. The chances are that by inoculating a trace of
that growth into a fresh nutritive material of the same kind as above,
we introduce only one species, because one species is only present
predominately. Growing different bacteria in different media in this
way, it is then found that they multiply or grow with different
degrees of rapidity. In a similar way success may be achieved by
adding to or subtracting from the nutritive medium certain sub-
stances.

In starting the cultivation of organisms, various methods are used
for inoculation. There is Koch’s method, by which a platinum wire,
heated and subsequently cooled, is used te introduce the trace ef the
organism into nutritive gelatin or other material, coutained in test-
tubes, plugged with sterile cotton-wool. In this process, the cotton-wool
plug is altogether removed. The test-tube containing the nutritive
material, which is solid, is inverted, and then the inoculation is per-
formed with the platinum wire by pushing the infected end of this wire
into the solid material ; the plug is replaced, the test-tube again put
upright and exposed to the suitable conditions favouring the growth of
the organism. In this process of inoculation, the chances of accidental
contamination are of course not avoided. In the summer months,
and in London, this method of inoculation is fellowed in a considerable
percentage of cases by undesired accidental contamination.

I myself prefer the plan in which accidental contamination is
reduced almost ‘to nil. I take a pointed capillary pipette, freshly
made, pass one end through the cotton-wool plug (previously drawn
up halfway.) down into the culture tube from which inoculation is to
be made, allow a drop to ascend into the capillary tube, withdraw it,
and pass it through the cotton-wool plug into the tube in which
growth is ‘to be started. In this way the chances of accidental
contamination, owing to the lifting out of the plug, are reduced
to the smallest dimensions, provided, of course, that the cotton-wool
plug is sterile. By the plate cultivation, and by cultivations in
test-tubes, the characters, both morphological and of the mode of
growth, can be studied and ascertained; and it will be found that
many different species have in this respect different characters. Some
grow in-spdts, others in flat patches, others as films, uniform or com-

PERKIN: THE CONSTITUTION OF UNDECYLENIC ACID. 205

posed of minute droplets; some grow best in the depth, others best
on the surface; sume liquefy gelatin, others do not liquefy it, and
so on.

Having, then, ascertained the characters of an organism, and
obtained it in pure cultivations, the organisms are then tested for their
activity by inoculating with them the suitable nutritive material. If
this organism, by multiplication in this material, is capable of start-
ing a specific chemical action in this new suitable material, then we
are bound tw say that it is this organism which is the cause of the specific
action. The whole chain of evidence must be complete in the sense
which I have indicated, in the sense in which it is considered as com-
plete by pathologists of the present day, and only then, and not till
then, will the statements put forward by chemists command that value
which they claim. Not till chemists come to look upon the matter in
the same light in which we look upon it, namely, to obtain the
organisms pure, to render nutritive material sterile, to be able to
produce with this pure organism the specific chemical activity you
wish to obtain—not till you have fulfilled all these conditions, can you
claim to have established the faet that a definite organism is the cause
of a definite chemical process.

XXIII.—On the Constitution of Undecylenic Acid, as indicated by its
Magnetic Rotation, and on the Magnetic Rotation, §c., of Mono- and
Di-allylacetic Acids, and af Ethyl Diallylmalonate.

By W. H. Perxiy, Ph.D., F.R.S.

Amonast the substances examined when first studying the subject of
the magnetic rotation of compounds was undecylenic acid, an un-
saturated acid of the formula Cy,H.Q,, discovered by Krafft (Ber., 10,
2035), as a product of the distillation of castor-oil under reduced
pressure. The measurements of this substance were not given in
the paper I had the honour of laying before the Society some time
since, because they did not appear to be consistent with those I
obtained for other unsaturated compounds. I was desirous of further
considering them, as I felt sure that a careful study of the subject
would clear up this apparent discrepancy.

Both the acid and its ether were examined. The acid was obtained
from Kahlbaum, and purified by rapid distillation under reduced
pressure. It boiled at 230—235° (corr.) at 130 mm. Krafft has

206 PERKIN: THE CONSTITUTION OF UNDECYLENIC ACID,

shown that it undergoes some amount of change, even when distilled
under reduced pressure, so that the specimen cannot be considered
quite pure. I find that this acid does not appear to undergo any
alteration when dissolved in alkali and treated with sodium amalgam
for two or three days. The acid examined was nearly white and
beautifully crystalline. Its density when in the fused state was
found to be as follows :—

Q5°
d= 091020,

45°
dj 0'89930.

The following numbers were obtained for its magnetic rotation :—

t. Sp. rotation. Mol. rotation.
26°5° 11153 12°534
26°5 11157 12°537
24°5 11174 12°541
24°5 11209 12°575

Average 25°5 11174 12°547

Ethyl Undecylenate-—This substance was obtained by saturating an
alcoholic solution of the acid with hydrochloric acid gas. The
etherification takes place easily, so that it is not necessary to keep
the product long in contact with the hydrochloric acid, which would
probably be unadvisable as it is an unsaturated compound. The new
ethereal salt was separated by the addition of water, and further
purified by solution in ether, washing with sodium carbonate, drying
over anhydrous potassium carbonate, and distilling. After fractioning
a few times, it was obtained as a colourless oil, boiling at 263°5—
265°5° (corr.).

Ethyl undecylenate is much more stable than the acid, and appears
to undergo but little change when distilled. It has a peculiar fruity
but not pleasant odour.

On analysis it gave the following numbers :—

0°1175 gram of substance gave 0°1200 gram of H,O and 0°3156
gram of CQ).
Found. Cale. for C3H 24 0x.
73°11
11°32

Dr. Gladstone has been good enough to measure the refractive
power of this substance, and obtained the following results :—

AS INDICATED BY ITS MAGNETIC ROTATION. 207

: Specific Specific Ref.
f. BA. PD. PH. refraction. dispersion. equiv. A.

23° «14393 14449 1°4611 0°4924 0°0244 90°60

He remarks that “ the refraction equivalent of this compound scarcely
decides as to the number of: carbon-atoms double-linked, but if we
suppose only one pair, the theoretical amount would be 89°4, and
therefore not much below the experimental, and the specific disper-
sion strongly supports that view.”

Density determinations of this substance gave the following
numbers :—

a
d5= 088271,

25° ones
d= 0°87658.

The following are the determinations of its magnetic rotation.

t. Sp. rotation. Mol. rotation.
24°5° 1°0855 14°589
25°0 1:0801 14-528
16°4 1:0857 14502
16°4 10856 14501

Average 20°6 10842 14530

If we now compare the molecular rotation of undecylenic acid with
that of its ethyl salt, allowing for the change of composition, we get
the following result :—

Undecylenic acid, 12°547 + 1023 x 2 .. 14570
Ethyl undecylenate

This corresponds very closely with the difference found in the case
of the fatty series, which is 0-056.

When these substances were examined, it was expected they would
give numbers having a series constant (Trans., 1884, 546) of 1°451,
the same as that of ethyl «-crotonate and oleate. In that case, the
molecular rotation of the ether would have been 14°750 instead of
14530, which has been found.

On referring to my previous paper (Trans., 1884, 561), it will
be seen that two sets of numbers were obtained, when comparing the
molecular rotations of unsaturated and saturated compounds. The
following is the table given :—

208 PERKIN: THE CONSTITUTION OF UNDECYLENIC ACID,

Ethyl a-crotonate
butyrate .........64.
oleate
stearate (calc.) 20°797
allyl malonate 11°281
»» propyl malonate 10°367
Allyl alcohol
Propy! alcohol

From this it is seen that allyl compounds give a smaller difference
in relation to the saturated compounds than the «-crotonic and oleic
compounds. ,

Now if the molecular rotations of undecylenic compounds be com.
pared with those of the corresponding saturated compounds, it will
be seen that numbers are obtained which differ only to a very slight
extent from those of allyl compounds.

Undecylenic acid merely 0-897
Undecylic acid 11-650
Ethyl undecylenate aaa 0890.
undecylate 13°640

”

These results indicate, therefore, that undecylenic acid is an allyl-
derivative. This being so, it was thought desirable to examine some
allyl-derivative of the fatty acids, which, if the above be true, should
be a homologue of undecylenic acid, and thus a stricter comparison
of the rotations could be made. For this purpose, allylacetic acid
was chosen, and whilst preparing this, it seemed to be of interest
to prepare diallylacetic acid, and examine its rotation also. The
following is a brief account of the way in which these compounds
were obtained.

Ethyl allylmalonate was first made in the usual manner, and frac-
tioned under reduced pressure, to remove unchanged ethyl malonate
and any ethyl diallylmalonate which might have been formed. The
ether was saponified with caustic potash and the product evaporated
on the water-bath until free from alcohol. The alkaline mass was
then dissolved, strongly acidified with hydrochloric acid, and the
allylmalonic acid thus set free taken up by repeated treatment with
ether free from alcohol. The ethereal solution was shaken with cal-
cium chloride to remove as much water as possible, decanted, and
the ether distilled off ; the acid was then dried on the water-bath, and
purified by crystallisation from benzene, which is a very convenient
solvent for this purpose. In this way, it was obtained in fine white
needles.

Ethyl! diallylmalonate was also prepared in the usual way from

AS INDICATED BY ITS MAGNETIC ROTATION. 209

ethyl allylmalonate. Some considerable quantity was made, which
was carefully fractioned several times under reduced pressure, and its
magnetic rotation and density again determined, as I was not quite
satisfied with my previous determinations (Trans., 1884, 539). The
boiling point of the product was 207-5—208'5° (corr.) under 260 mm.
The density determinations gave—

65°.
a 1:00620.

15° 4.
a 0-99940.
a> 0-90252.
These numbers are not very different from those of the specimen
previously examined, which were for—
15° 25° 2.
d= 099997, and for a5 0°99301.
The following are the numbers obtained for the magnetic rotation
of this new preparation :—

t. Sp. rotation. Mol. rotation.
13°5° 1:1277 15°030
13°5 1°1265 15013
140 1:1274 15031

Average 13:7 11272 15°025

The previous number obtained for the molecular rotation was
14°998, and therefore only differs from this mew one by 0°027.

To obtain the diallylmalonic acid, the ether was treated with
caustic alkali, and all the aleohol driven off from the alkaline product
on the water-bath ; it was then dissolved and acidified strongly with
hydrochloric acid, when the diallylmalonic acid separated as a
magma of fine crystals. After standing for some time, these were
collected on a calico filter and well pressed from the saline solution ;
the latter being treated with ether to recover any acid remaining in
solution. The crude acid was crystallised two or three times from
water, and in this way was obtained in beautiful large prisms. Pro-
fessor Haushofer has been good enough to measure these crystals, and
gives the following account of them.

“ Crystalline system, Rhombic.”

Transparent brilliant crystals of the combinations coPco(100) = a.
coPco(010) = b. P(111)=o0. Usually prismatically developed by

210. PERKIN: THE CONSTITUTION OF UNDECYLENIC: ACID,

the predominance of the vertical surfaces a and b. As 4 ¥ulé the
surface a slightly predominates; frequently, however, both surfaces

are equally developed, and then the crystals have a distinct tetra-
gonal habitus, about which the measurements of the angles afford no
evidence, as there is but little difference between the axes a and b.
Cleavage towards b, but not well marked. The surfaces a and b are
usually uneven and not adapted for measurement; there is so great
variation in the angles that the determination of the crystalline
system rests on the optical properties. Occasionally some indication
of the primary brachydome and of the base may be observed :—

Measured. Calculated.
0 : 0 (111) (111) = 108° 54’ — — (Brachydiagonal polar
angle.)
0:0 (111) (111) = 108 36 — — (Macrodiagonal polar
angle.)
0:0 (111) (111) = 110 53 110° 56’ (Basal angle.)

“A plate cut parallel to the base (which is only effected with difficulty
owing to the brittleness of the crystal) shows an interference image
in convergent polarised light like a uniaxial crystal; but on turning
the plate on the vertical axis the dark arms of the cross separate into
two hyperbolas.

“It is difficult to decide how the axial planes lie, as that arm of the
cross which is parallel to the macrodiagonal is considerably the
stronger. I am inclined to believe that the optical axes lie in the
plane of the macrodiagonal. No determination of the relative dis-
persion or of the optical characters of the substance could be made on
account of the weakness of the interference image.”

AS INDICATED BY ITS MAGNETIC ROTATION. 211

The allyl- and diallyl-acetic acids were obtained by heating the
corresponding malonic acids—after being thoroughly dried—in a
retort with a reflux’ condenser until carbonic anhydride ceased to
be given off. They were then distilled; their formation takes place
very rapidly.

Allylacetic acid thus obtained boiled at 185—188° (corr.), the small
variation in its boiling point being probably due to the presence ofa
minute quantity of the diallyl acid. It has a sharp, acid odour, not
unlike that of valeric acid, but with a slight train oil character. It
does not solidify when cooled in ice and hydrochloric acid. It floats
on water as an oil, but is soluble to some extent in this fluid.

The density determinations gave the following results :—

12° ...
dy 0°98656.

15° 4.
a 0 98416.

_ eres
d= 0°97670.

The following numbers were obtained for its magnetic rotation :—

t. Sp. rotation. Mol. rotation.
12°5° 1°1429 6°439
12°5 1°1403 6°424
140 Pts 1°1369 6°413
140 1°1380 6°419
14°6 1:1398 6°432
14°6 1:1389 6°427
14°6 11403 6°435
146 11382 6°425

Average 13°9 1:1394 6°426

= 2
E.
<j
=
£2. |
ct
“4
&
te

&
3
t
t

-
“va

The diallylacetic acid obtained as above boiled at 227—227°5° (corr. ).
Ii has an odour similar to that of train oil, but somewhat sharper.
When cooled with ice and hydrochloric acid, it thickens, but does not
solidify: it is only slightly soluble in water. Determinations of its
density gave the following numbers :—

12°.
d= 0-95756,

15°
oan oF 7
d5=- 095547,
25°

“094913.
d= 094913

212 PERKIN: THE CONSTITUTION OF UNDECYLENIO ACID,

The measurements obtained for its magnetic rotation were as
follows :—

t. Sp. rotation. Mol. rotation.
18°5° 1:2681 10-346
18°5 12682 10347
18°5 1:2694 10°357
18°5 1:2692 10-356
13:0 1:2708 10-329
13°5 1:2720 © 10-343
140 1:2699 10-329

Average 16°4 1:2697 10°344

If from the molecular rotation obtained for allylacetic acid we sub-
tract that of the corresponding saturated compound, valeric acid, we
get the difference due to the influence of the allyl-group—

Allylacetic acid
Valeric acid .

If we do the same with diallylacetic acid, and subtract from it the
molecular rotation of octylic acid, we get—

Diallylacetic acid
Octylic acid

This number is the difference due to the influence of two allyl-groups,
but it will be seen that it is less than twice that of the first replace-
ment; if we subtract the one from the other we get—

Two allyl-groups ..... ececccece °
- Ist allyl-group

2nd ” ”

This is quite consistent with previous observations on successive
replacements of all kinds. In the case of ethyl diallylmalonate, the
influence of this second replacement cannot be found because ethyl
dipropylmalonate has not yet been examined, and the malonic series
has not been sufficiently studied to calculate this. There is very little
doubt, however, that it is nearly the same as in the above, and if
so we then find that ethyl dipropylmalonate should have a rotation of
about 13246, which is very probable judging from other results.

In reference to undecylenic acid, however, the interest attaches to

AS INDICATED BY ITS MAGNETIC ROTATION. 213

monallyl compounds. In my previous experiments, I found that the
difference between an allyl and a saturated compound was 0°914.
This is further confirmed by the examination of allylacetic acid,
which gives 0°918, a number a little higher, but the difference is so
very small that the two may be regarded as identical. This result
therefore strengthens the inference that undecylenic acid is an allyl
fatty acid. Still the numbers which are already given in this paper
for this substance and its ethyl salt being part of some of the earlier
work on the magnetic rotation, to which a small correction, since
discovered, should have been applied, it was thought desirable to re-
measure either the acid or the ethyl salt; the latter was selected as
being the purer product, and, being liquid at the ordinary tempera-
ture, is more convenient for examination. The new results were as
follows :—

t. Sp. rotation. Mol. rotation.
13°6° 10898 14°528
13°4 10894 14554
13°8 1:0910 14°526
13'8 1:0908 14543
14°6 10917 14564

Average 13°6 1:0905 14°543

Now if from this molecular rotation we subtract that of ethyl
undecylate we get—

Ethyl undecylenate
» undecylate

and it will be seen that this difference is nearly identical with that of
the allyl compounds, being only 0-011 less, which, considering the
high molecular weight of the substance and comparatively low specific
rotation, is as close as could be expected.

From the foregoing results, it is seen that undecylenic acid and its
ethyl salt give magnetic rotations which are peculiar to monallyl-
derivatives of the fatty acid and other series, being lower than any
other unsaturated compounds yet examined. This remark refers to
the bibasic acids, citraconic, itaconic, and maleic, as well as those given
in my previous paper, and thus indicates that the acid in question is an
allyl-derivative of a fatty acid. If this be so, and there is no reason
for doubting it so far as can be seen, the next question is as to what
is the position of the allyl-group, whether this acid be, for example,
an allyl-hexyl-acetic acid or an allyl-octylic acid—

VOL. XLIX. Q

214 PERKIN: THE CONSTITUTION OF UNDECYLENIC ACID.

Cin>CH-COO8, or C;HsC;H,,,COOH.

The answer to this is given by the oxidation product it yields when
treated with nitric acid. Krafft has shown that under these circum-
stances it yields sebacic acid; now for this to be formed it must be
an allyloctylic acid, which may be written thus :—

CH, : CH:CH,*(CH,),,COOH.
The oxidation would remove the CH, and convert the CH next to it
into COOH, yielding
COOH-(CH.,),-;COOH,

in the same way as its homologue, allylacetic acid, under similar
circumstances, gives succinic acid.

There is one reaction of undecylenic acid which is very remarkable,
and that is, that when it is fused with caustic alkali it yields nonylic
acid; this reaction is difficult to explain unless sebacic acid is first
formed and then gives up carbonic anhydride, thus :—

COOH-(CH,),COOH = CH,°(CH:),,COOH + CO.,

but, judging from small experiments made in this direction, this does

not seem to be the case.
If undecylenic acid had the constitution

CH,(CH,);CH : CH-COOH,

the formation of nonylic acid under those circumstances would be in
accordance with experience, but this cannot be the case, otherwise it
would have a munch higher molecular rotation, and moreover it would
not yield sebacic acid on oxidation with nitric acid. It therefore
appears that undecylenic acid is allyloctylic acid or, which is the
same thing, vinylnonylic acid,
CH, : CH:(CH,),,COOH.
I hope shortly to bring before the Society an account of the mag-

netic rotations of the unsaturated bibasic avids and their so-called
physical isomerides.

XXIV.—On the Condition of Silicon in Pig Iron.

By A. E. Jorpan (Student in the Chemical Laboratories, Mason
College, Birmingham) and THomas TuRNER.

Siticon, like carbon, is stated to exist in three forms, the amorphous
variety corresponding to charcoal, the graphitic to graphite, and the
adamantine to the diamond. Amorphous silicon is distinguished by
its solubility in hydrofluoric acid. The chemical properties of graph-
itic and adamantine silicon are said to be the same, both varieties
being insoluble in hydrofluoric acid, and unaffected by heating in
oxygen, though soluble in a mixture of nitric and hydrofluoric acids.
The graphitic variety forms thin hexagonal plates, while the adaman-
tine modification crystallises in octahedra. It has been affirmed by
Professor W. H. Miller that the plates of so-called graphitic silicon
are modifications of the octahedral form, and it is therefore probable
that only two really distinct varieties are known.

As silicon is a constant constituent of pig iron, and as it may
readily be obtained in a crystalline form by separation from solution
in various metals, it seems natural to expect that silicon, like carbon,
would separate from cast iron in the graphitic condition. It is
generally so stated in metallurgical books, and a number of workers
have at various times confirmed this view.

Thus Mr. J. A. Phillips (“ Metallurgy,” p. 115) says, “Cast iron
often contains silicon in the graphitic state,” while Dr. Percy,
speaking of the insoluble residue left by the action of diluted acids on
grey cast iron, says (“Iron and Steel,” p. 145), “‘ There is reason to
believe that crystallised silicon may also not unfrequently be present,”
and adds, “ Mr. T. H. Henry informed me that he found crystallised
silicon amongst the graphitic scales obtained by the action of hydro-
chloric acid on pig iron.” In Wagner’s Jahresbericht (1857, iii,
8) a reference is given to some analyses by Wohler, who is stated
to have found crystallised silicon in cast iron. Professor Robert
Richter, of Leoben, also observed crystallised silicon in cast iron
(Berg und Huttenman, Jahrbuch, 1862, 289). A specimen of finely
crystallised cast iron was extracted with dilute hydrochloric acid so
long as anything dissolved, the residue was washed, dried, and ignited
ina stream of oxygen. It was then extracted with hydrochloric acid
to remove oxide of iron, washed, and dried. It now contained thin
graphitic-looking plates, with a metallic lustre, which were not
affected by further heating in oxygen. These plates were readily
oxidised when heated with potassium nitrate and sodium carbonate,

Q 2

216 JORDAN AND TURNER: THE CONDITION OF

yielding some oxide of iron and silica. Richter concluded from this
that these plates were identical with Deville’s crystallised silicon, then
newly prepared, and appears to have regarded the oxide of iron
merely as an accidental impurity.

An important paper was read by Mr. Snelus at the Merthyr meeting
of the Iron and Steel Institute, 1870, and was afterwards published
in the first volume of the Proceedings of the Institute (p. 28). In
this paper, conclusions were arrived at which are entirely different
from those previously mentioned. Thus, while Snelus was able by
means of a magnet, or by careful sifting, to more or less completely
separate graphitic carbon from cast iron, he was unable in this way to
separate silicon. On the contrary, the proportion of silicon present
in the iron was actually increased, probably owing to the separation
of carbon by the methods employed ; and after a number of careful
experiments on pig iron containing a considerable proportion of
silicon, Snelus concluded that silicon “ must have been in solution or
combination in the iron, and that it is at least an exceptional case if
it is found in the free state ” (p. 35).

The question was attacked in an entirely different manner by
Mr. Morton (Chem. News, 29, 107), who showed that the silicon
in a silicious pig iron (4°6 per cent. Si) was wholly converted into
silica by heating it in sealed tubes, either with Nordhausen sulphuric
acid or with iodine, in an atmosphere of carbon dioxide; and although
graphitoidal silicon was sought for it was not found in these experi-
ments.

Some interesting observations have also been made by Dr. Tilden
(Proceedings Birmingham Philosophical Society, 3, 203), who, when
examining a silicious Staffordshire pig iron, obtained a quantity of
black scales in the residue after treatment with dilute hydrochloric
acid which did not burn when heated in a muffle, but which were
proved to contain iron, and from which amorphous silicon was
separated, as a brown powder, by protracted boiling with hydrochloric
acid. Dr. Tilden suggested that by prolonged boiling, the residue
might have been caused to yield up the whole of its iron, and that
possibly in those cases where graphitic silicon was supposed to have
been observed, the substance was in reality a silicide of iron or man-
ganese, resembling the compound described.

Our own experiments were of three kinds, including—

1. A magnetic examination of certain specimens which appeared
most likely to contain crystallised silicon.

2. An examination of the residue left on prolonged treatment with

dilute hydrochloric acid.
3 Some experiments on graphitic matter separated from cast iron.

SILICON IN PIG IRON.

Magnetic Examination.

Five specimens of cast iron were operated upon in this manner, but
as the results in each case are similar, so far as the absence of
graphitic silicon is concerned, it will perhaps be sufficient if we de-
scribe the experiment which we consider most conclusive.

35 grams of silicon pig, containing 9°8 per cent. of silicon, was
taken after crushing and passing through a 60 sieve. It was laid
upon a clean piece of paper, and then everything attracted by means
of a magnet was removed on to a similar sheet of paper, by drawing
the magnet repeatedly through the mass. This was again and again
repeated so long as any perceptible residue was left. These residues
were mixed together and further treated by the magnet, and at length
a small residue was obtained which appeared to be quite indifferent to
magnetic influence. The weight of this residue was 0°11 gram, or
0°32 per cent. of the original material. When examined under the
lens, it was seen to contain some black graphitic particles, which
entirely burned off on afterwards heating in the muffle, and were
evidently graphitic carbon. Some slightly coloured, sometimes
angular, and at other times rounded fragments were present; they
appeared to be nearly transparent, and were probably sand, either
accidentally introduced or perhaps mechanically intermixed during
the running of the pig. The greater part of the residue which did
not burn off in the muffle, however, was nearly white, quite opaque,
and generally in rounded nodules ; this was evidently slag, of which a
similar specimen of iron has already been shown to contain 0°76 per
cent. (Turner, Trans., 45, 265). Now since the graphite in this
iron is only present to the extent of about 1 per cent., and slag to
an even smaller extent, it is scarcely conceivable that graphitic silicon
could exist in any quantity in the iron without being separated, at
least to some extent, along with the graphite and slaggy matter. And,
certainly, if graphitic silicon does occur in cast iron at all, silicon pig
would appear to be the most probable material in which it should be
found.

Examination of Insoluble Residues.

For these experiments, five specimens of iron were examined, in
which the proportions of silicon and graphitic carbon varied consider-
ably. The method adopted in their examination was proposed to us
by Dr. Tilden, to whom we are indebted for several suggestions in
connection with this part of the work.

1. South Staffordshire White Iron—The metal was crushed in a
steel mortar, and passed through a sieve; 40 grams was taken and
boiled repeatedly with dilute hydrochloric acid (about 1 of strong

218 JORDAN AND TURNER: THE CONDITION OF

acid with from 2 to 3 of water) until no more iron was extracted, as
indicated by testing the solution with potassium ferrocyanide. After
being thoroughly washed with hot water, the residue was dried and
heated in a muffle, the temperature of which was at about the point
necessary for the cupellation of silver. The residue burned quite
white, and was entirely soluble in hydrofluoric acid, and could there-
fore have contained no graphitic silicon. In order to be certain that
heating in the muffle was not capable of oxidising graphitic silicon, we
placed a specimen of the pure substance in the muffle under exactly
the same circumstances as the residues to be examined, and after a
lengthened exposure the specimen of graphitic silicon was apparently
unaltered.

2. No. 3. Hematite Pig.—This iron contained 2°6 per cent. of silicon,
and was used in the form of borings, 40 grams of which was taken
for the experiment. It was repeatedly extracted with dilute hydro-
chloric acid as before, the residue thoroughly washed, dried, and
ignited in the muffle. The residue was quite white, and entirely dis-
solved in hydrofluoric acid. It could therefore have contained no
crystallised silicon.

3. No. 2. South Staffordshire Grey Iron.—In this case, 40 grams was
heated with hydrochloric acid, &c., exactly as before. But on burning
off the carbon a residue was left, slightly brown in colour, and on
extraction with hydrofluoric acid a brown residue remained. It was
very small in quantity, and resembled oxide of iron in appearance.
Though insoluble in hydrochloric acid, it yielded a portion of iron
when evaporated with hydrofluoric and hydrochloric acids alternately.
When dried and examined under the microscope, nothing resembling
graphite was observed in the residue, which was probably silicide of
iron partly oxidised by roasting in the mufile.

4, No.1. Middlesbrough Grey Pig (Clarence Works).—In this case,
the same quantity was taken as before, the methods adopted being
also similar. The results obtained exactly resembled those of the first
two specimens.

5. Govan Silicon Pig.—This specimen contained 10°3 per cent. of
silicon. It was crushed in a steel mortar, and sieved to ensure
uniformity in size of the fragments, 40 grams being taken for the
experiment. This was boiled repeatedly for many hours with dilute
hydrochloric acid, until no more iron was dissolved by the liquid.
The residue, after washing and drying at 100°, was of considerable
quantity and grey in colour. A portion of this was roasted in the
muffle, when a slightly brown residue was left. On extracting this
with hydrofluoric acid, brilliant scales were left which much resembled
graphitic silicon in appearance; but on roasting these scales for a
considerable time in the muffle they gradually lost their lustre,

SILICON IN PIG IRON.

becoming dark brown and much like oxide of iron in aspect. This
residue was quite insoluble in hydrochloric acid even on boiling, but
on evaporating repeatedly with hydrofluoric acid it then yielded a
small quantity of iron when again extracted with hydrochloric acid.
This brown substance was, however, soluble when heated for some
time with hydrochloric acid in a sealed tube at 140°, and on quali-
tative examination proved to contain titanium in small quantity.
There is no doubt, therefore, that the scales which remained, after
roasting and extraction with hydrofluoric acid, consisted almost
entirely of difficultly decomposable silicide of iron, while no crystal-
lised silicon was present. It will be remembered that in Richter’s
experiments, previously mentioned, a small quantity of oxide of iron
was obtained from the graphitic scales, in addition to silica. They
may therefore very possibly have resembled those obtained in our own
experiments.

A portion of the residue, before roasting, and after drying at 100°,
was next examined. It was mixed with a little solid potash in a tube,
and arrangements made for collecting any gas which might be
evolved. On melting the potash by applying a gentle heat, a violent
action ensued and some hydrogen was evolved. This has been con-
sidered by some writers an evidence of the existence of free silicon ;
but it has been stated by Allen (Chem. News, 29, 91; 40, 65) that
the solution of silicious pig iron in dilute hydrochloric or sulphuric
acid gives rise to the production of Buff and Wohler’s “ leucone ”
(Watts’ Dict., 5, 274, Miller, Part II, p. 280, 1878), and this is
probably the correct explanation of the action. Leucone, according
to Wohler, has the formula 3Si0,2H,0, and may be regarded as a
hydrate of the monoxide of silicon ; but Friedel and Ladenburg regard
it as silicon formanhydride (Si,H,0;). The residue to which we have
veferred dissolved readily in weak warm hydrofluoric acid, evolving
hydrogen with effervescence ; it also decolorised acidified solution of
potassium permanganate; and when strongly heated in a test-tube,
partly closed at the top by the finger, a slight explosion took place,
caused by the evolved hydrogen. In these characters, as wel! as in its
permanency at 100°, and the evolution of hydrogen with potash, it
exactly resembles leucone. it is probable, however, that only a small
portion of the residue consisted of this substance, as it is readily
oxidised in the presence of ferric salts, and hence would be gradually
destroyed in the process of solution. But the fact of its formation
negatives any evidence in favour of the existence of crystalline silicon
in cast iron, founded upon the evolution of hydrogen on treatment of
these residues with molten potash. It will be remembered we have
previously shown that this residue does not contain graphitic silicon,
though it does contain scales which resemble that substance.

220 JORDAN AND TURNER: THE CONDITION OF

In the foregoing experiments on residues from cast iron, we have
examined specimens which were intended to be as far as possible
representative, and in no case have we met with any evidence of the
existence of crystalline silicon. We have, however, observed facts
closely resembling those adduced in support of its occurrence in cast
iron, and which may help in some measure to account for mistakes in
the case of some earlier observers. For it must be remembered that
about 25 years ago, when the work of Deville and Wohler was
being performed or was quite new, crystalline silicon was often sup-
posed to have been observed, while now, when analyses are much
more frequent, its occurrence is seldom or never recorded.

Examination of Graphite.

Since the amount of graphitic carbon is generally greatest when
carbon and. silicon are present in nearly equal proportions, it appeared
probable that graphitic matter separated from cast iron would contain
crystallised silicon, if it ever separated from cast iron in this condi-
tion. We therefore examined two very fine specimens, one of graphite
and the other of kish, for both of which we are indebted to the kind-
ness of Dr. Percy.

1. Graphite from Dowlais—This specimen was forwarded to Dr.
Percy from Mr. Child, of Dowlais, one of his former pupils. It was
examined by Dr. Percy, and is mentioned in his “Iron and Steel,”
p- 145. After being digested with hydrochloric acid, it was washed,
dried, and put into molten potash in a gold crucible. ‘Much efferves-
cence took place, due to the evolution of gas, which ignited at the
surface just like hydrogen. When native graphite was similarly
treated there was not the slightest effervescence.’ Dr. Percy re-
garded this action as probably caused by free silicon.

In talking of this experiment a short time since with one of us,
Dr. Percy in the kindest manner offered to place the whole of this
specimen which he had left in his collection at our disposal if neces-
sary, so that the correctness of the above inference might be ascer-
tained. For this purpose, a quantity was taken and digested with
strong hydrochloric acid, and afterwards with a 5 per cent. solution
of potash, neither of which reagents appeared to be capable of ex-
tracting anything from the specimen. It was then dried and heated
in the muffle. Part burned off very rapidly, while another portion
required much longer for its combustion, but ultimately only a very
small brown ash remained. This dissolved in boiling hydrochloric acid,
leaving only two or three minute white specks, while the solution con-
tained a small quantity of iron. The specimen would appear, therefore,
to be almost entirely free from silicon in any form, and to be remark-

SILICON IN PIG IRON. 221

ably pure graphitic carbon. It is quite conceivable from the known
variations in the igniting point of different varieties of carbon, that a
portion of this pure form of graphite might evolve a combustible gas
in contact with fused potash, while the less combustible part and
native graphite would remain entirely unaffected at the temperature
employed. This is probably the explanation of the effect observed by
Dr. Percy.

2. Kish from Middlesbrough.—F¥or this fine specimen, we are also
indebted to Dr. Percy. It was part of a larger quantity forwarded
to him from Clarence Works. It was repeatedly extracted with
hydrochloric acid, which removed a considerable quantity of iron and
also some manganese. The residue was extracted with 5 per cent.
solution of potash and some silica removed. There were left behind
some fine plates of graphite, and what was probably a few grains of
sand with which the kish had been contaminated. On drying and
burning in the muffle, the residue was light brown, but became quite
white on prolonged digestion with hydrochloric acid. This specimen
was therefore free from crystallised silicon.

In addition to the facts already considered, there is an argument in
support of the existence of crystallised silicon in cast iron, which was
mentioned by Sir I. Lowthian Bell in the discussion which followed
the paper by Mr. Snelus, to which previous reference has been made.
Under certain circumstances, pig iron is observed to be covered with
a peculiar soft downy coating on its upper surface, resembling wool
in appearance, but consisting of silica. It has been suggested that
this is produced by the oxidation of silicon exuded by the iron in
solidification, and that probably this variety of iron contains graphitic
silicon. But such an argument can by no means be taken as proof of
the fact; and we may so far anticipate a future communication as to
state that an exactly similar covering of silica has been observed by
one of us, in a case where the most careful examination failed to yield
any evidence of the presence of graphitic silicon.

Weare aware that a considerable accumulation of negative evidence
is necessary before it can be considered absolutely conclusive. At
the same time we feel that in this matter a tolerably strong case has
been made out. Thus, while it is not possible to separate silicon from
iron either by sifting or by a magnet, and it cannot be detected in
the residue left by the action of a diluted acid or in graphitic matter,
it has been obtained in the amorphous form by Dr. Tilden from scaly
silicide of iron. It is probable, therefore, that silicon exists in cast
iron in the form of a silicide (or silicides) which is dissolved when
fluid in the excess of iron present. The silicide itself is attacked by
acids only with considerable difficulty, but when present in small
quantity, being more finely divided, it is easily decomposed. On the

222 WILLIAMS: REACTIONS SUPPOSED TO YIELD

other hand, evidence in favour of the existence of crystallised silicon
in cast iron is at present very unsatisfactory, and would require to be
considerably strengthened before it could be accepted.

XXV.—Reactions supposed to yield Nitroxyl or Nitryl Chloride.

By W. Coxtinewoop Wix.iams, B.Sc., Tangye Scholar in the Mason
Science College, Birmingham.

In all the larger text-books that I have examined, it is stated that
nitryl chloride not only exists, but is readily prepared.

These statements are based on the researches of Odet and Vignon
(Compt. rend., 69, 1142, and 70, 96), Williamson (Proc. Roy. Soc., 7,
15), and Hasenbach (J. pr. Chem. [2], 4, 1).

When these papers are examined, however, it is at once obvious
that the definite and plausible statements of the text-books are based
on very imperfect evidence. Odet and Vignon, in the papers cited,
were working in the hope of elucidating the process introduced by
Deville for the preparation of nitrogen pentoxide. The formation of
the latter by the action of chlorine on silver nitrate, they explain by
the equations—

bo + Cl = AgCl + NO,Cl + 0

_ NO; .
- ho + NO.Cl = No: yO + AgCl

and in order to test the correctness of these equations, they resolved
to bring nitroxy] chloride into contact with nitrate of silver.

It is therefore with some surprise that one learns that in order to
prepare the nitroxyl chloride they have recourse to a reaction in
which, if formed, it will on their own hypothesis be immediately
decomposed, viz., the action of phosphorus oxychloride on excess of
lead or silver nitrate. It is stated that the product boiled at 5°, and
it is described as of a pale-yellow colour.

In their second investigation, in which they endeavoured to obtain
the nitroxyl chloride formed as an intermediate product (according to
their view) in Deville’s process, they say that they were five hours in
collecting 2 c.c. of liquid, which they assume to be nitroxyl chloride.
It boiled in the hand, and was yellowish-brown in colour.

In neither paper is there any account of a quantitative examination,
nor did they take any steps to identify the compound.

NITROXYL OR NITRYL CHLORIDE. 223

Thus far Odet and Vignon. Turning to Williamson’s paper (loc.
cit.), we find that he poured sulphuryl chlorhydrin, SO,Cl-OH, on
dry fused nitre, and found that a gas smelling of aqua regia was
evolved, giving hydrochloric and nitric acids when passed into cold
water. This, he remarks, is doubtless nitryl chloride.

Hasenbach passed a mixture of chlorine (in excess) and nitric
peroxide through a strongly heated combustion-tube. Since no tem-
perature is stated, I think one is justified in supposing that the tube
was red-hot. As to the product, he merely says that it was “ pos-
sessed of the properties of NO,Cl.” After one rectification the liquid
was considered pure and contained 44 per cent. of chlorine, the
theoretical quantity for NO,Cl being 43°5 per cent.

This coincidence I regard as purely accidental; for in the first
place, the analysis of only one specimen is given, and secondly, in my
own experiments, when working under like conditions, I have never
obtained a liquid with even approximately a constant boiling point,
or which could be considered pure after one redistillation.

The large proportion of chlorine is readily accounted for by the
fact, as I shall show, that nitrosyl chloride, NOCI (54 per cent. Cl), is
® principal product, and further, that chlorine dissolves to a large
extent both in NOC and in N,Q,.

It is also worthy of remark that in endeavouring to prepare NO,Br
by the direct union of NO, and Br, Hasenbach found only 34 per cent.
Br in the product, as against 63 per cent. theory. NO,I could not be
prepared.

On the other hand, Armstrong in 1873 found that when acetic
chloride acts on a nitrate, the components of nitryl chloride are formed,
but not the substance itself. Peroxide of nitrogen and chlorine are
evolved, while a certain quantity of acetic anhydride is produced
simultaneously (Chem. Soc. J., 1873, 26, 683).

From nitrites on the contrary, nitrosyl chloride was readily ob-
tained by similar treatment.

Quite recently, however, reactions of the same class have been re-
examined by Lachowicz (Ber., 17, 1281; 18, 2990). He does not
seem to be aware of Armstrong’s experiments, for he makes no men-
tion of them, and states that from silver nitrate and an acid chlor-
anhydride, the acid anhydride is obtained. By this reaction, he
prepared acetic, benzoic, and phthalic anhydrides in theoretical
quantity. The gaseous products are peroxide of nitrogen and oxygen.
Here he differs both from Armstrong and myself, but in any case
the reaction does not avail for the preparation of nitryl chloride.
Lachowicz gives as the typical equation for this decomposition,
2AgNO, + 2X:COCI] = 2AgCl + [X:CO-],.0 + NO, + O, where X
is a hydrocarbon radicle.

224 WILLIAMS: REACTIONS SUPPOSED TO YIELD

Judging, therefore, by the light of previous research, it appears
that the evidence for the existence of nitryl chloride is by no meams
complete.

With the object of further investigating this question, and of pre-
paring nitryl chloride for use as a reagent, I have made a number of
experiments such as might be expected to yield the substance. I
shall now proceed to describe these.

1. Action of Phosphorus Oxychloride on Lead Nitrate in Excess.

This reaction was discussed some time ago by Mills (Phil. Mag. [4],
40, 134). Oxychloride of phosphorus contained in a tap funnel was
allowed to drop on to pure dry lead nitrate in a distilling flask,
connected with a spiral condenser cooled by a freezing mixture of ice
and salt, the flask being surrounded by warm water to promote
reaction.

This process fails utterly as a source of nitryl chloride. No liquid
product was obtained in any quantity, and throughout the experiment
a gas having the smell and bleaching properties of chlorine escaped.
The vapours in the flask had a reddish colour, resembling that of
dilute nitrogen peroxide, and after the experiment had gone on for
some hours, a few drops of a yellowish-red liquid were collected, but
the quantity was far too small to be of any use.

2. Action of pure Nitric Acid on Phosphorus Oxychloride in Excess.

The same apparatus was used, the POC], being now placed in the
flask. The reaction seemed at first more promising than the last, as
more liquid distillate was obtained. On examination, however, it was
found that all distillates contained POCI;, even though distilled at
20° to 20°, and it therefore seemed likely that the nitryl chloride, if
formed in this way, combined with the excess of POC).

3. Action of Nitrie Acid on Phosphorus Oxychloride in Equivalent
Proportions.

The materials were heated at 100° for about two hours in sealed
tubes. The contents separated into two layers, the upper deep red
and mobile, the lower yellow and viscid. The latter consisted doubt-
less of phosphoric acid. On opening the tubes, a very volatile vapour
having the odour of aqua regia escaped, which did not condense at
—19°. A portion of this vapour was collected over strong sulphuric
acid, and after absorption of the hydrochloric acid gas by water, was
found to be chlorine.

It is probable that through dehydration of the orthophosphoric

NITROXYL OR NITRYL CHLORIDE. 225

acid, and consequent formation of hydrochloric acid, the aqua regia
reaction took place, with production of chlorine and nitrosyl chloride.
On account of this complication, the reaction was not further studied,
but, after the experiments of Williamson with sulphuryl chlorhydrin
had been repeated, it was decided to use sulphuryl dichloride and
nitric acid, or potassium nitrate, in which case dehydration of the
resulting acid was not probable. I should mention that the reaction
between nitric acid and phosphorus oxychloride takes place, to all
appearance completely, in the cold after long standing. In this
respect, the action of phosphorus oxychloride is in marked contrast
with that of sulphury] dichloride.

4, Action of Sulphuric Chlorhydrin, SO.(OH)Cl, on Potassium Nitrate
in Equivalent Proportions.

When a small quantity of the dry fused nitrate is placed in a test-
tube and sulphuric chlorhydrin is added to it, the mixture becomes hot
and evolves chlorine. Subsequently, on applying heat, nitric peroxide
is given off in large quantity. No other substance save these two
could be detected, even when considerable quantities of the materials
were employed in equivalent proportions, and the evolved vapours
passed through a condenser cooled to — 18°. At first, chlorine alone

escaped and no liquid condensed, but on the application of heat to the
flask red vapours were formed, which on passing into the condenser
gave a small quantity of a deep red liquid now perfectly familiar to
me as a solution of chlorine in nitrogen peroxide. The issuing gas on
being passed into hot water always gave chlorine. With cold water, of
course hydrochloric acid is produced, due to the oxidation by the
chlorine of the nitrous acid formed by the solution of the peroxide in
water.

The “gas smelling of aqua regia” is therefore chlorine together
with a little nitrogen peroxide, and it is noteworthy that the evolution
of the chlorine before the peroxide was also observed by Armstrong
when potassium nitrate is treated with acetic chloride.

Possibly the nitric peroxide first formed combines with unattacked
chlorhydrin, forming a nitrosyl compound, in this way :—

(i.) 280,(0H)Cl + 2KO-NO, = 2S0,(0H)(OK) + N,0, + Ch,

(ii.) SO,(0H)Cl + NO-NO; = SO({O-NO)Cl + HNO,,

while the chlorine escapes. When heated, such a compound might
easily decompose in presence of excess of potassium nitrate, forming
chlorine and nitric peroxide, the latter being now in excess :—

(iii.) 280,(0-NO)Cl + 4NO,-OK = 3N,0, + Cl + 2S80,(OK):.

226 WILLIAMS: REACTIONS SUPPOSED TO YIELD

5. Activn of Sulphuryl Dichloride, SO,Cl,, on Potassium Nitrate.

This experiment was tried as a crucial test of this class of reaction.

Sulphuryl dichloride was prepared by heating the chlorhydrin in
sealed tubes for some days at 180°, the liquid being then distilled on
a water-bath. The product, which contained 98 per cent. SO.Cl., was
sealed up in a tube with an equivalent quantity of potassium nitrate
(upon which it could not be made to act in an open tube), and heated at
100° for some hours. A certain amount of red gas was formed, and
on opening the tube chlorine was collected over hot water. This does
not leave much room for doubt that the reaction is—

SO.Cl, + 2KO-NO, = N.0, + Cl, + S0,(OK)>.

The reaction takes place with great difficulty, and is very incom-
plete.

Towards absolute nitric acid, sulphuryl dichloride exhibited still
greater inertness.

From such experiments as I have described, I infer that generally
when a nitrate is acted on by an acid chloride—conditions under
which we might reasonably expect nitroxyl chloride to be formed—we
do not obtain this substance, but, instead of it, its component parts,
nitrogen peroxide and chlorine.

6. Action of Chlorine on Nitric Peroxide.

I now pass to the consideration of the method of direct union
suggested by Hasenbach.

Three principal series of experiments were performed under this
head: in the first, which I shall call series A, the chlorine and
peroxide of nitrogen were passed through a red-hot tube, in the other
two, the components were heated at a temperature not exceeding 150°.
The difference between series B and C will be explained later.

A. In these experiments, the chlorine, generated from manganese
dioxide (in lumps) and hydrochloric acid, washed with water, and
dried by sulphuric acid, was bubbled through liquid peroxide of
nitrogen, contained in a wide test-tube which could be warmed if
necessary. The vapours were thoroughly mixed in a dry Woulff’s
bottle and then entered the combustion tube, which was packed with
small pieces of clay pipe and heated to redness. Condensation was
effected in a spiral condenser cooled by ice and salt, the receiver, a small
Wiirtz flask, being also well cooled in a similar way. The uncon-
densed gas passed away by the delivery tube of the latter. A closed
receiver of this kind was found to be absolutely necessary, as many

NITROXYL OR NITRYL CHLORIDE. 227

experiments failed owing to the rapid action of the moisture of the
air on the product.

Now, since NO, dissociates at a red heat into NO and oxygen, it
was likely that NOCI would be a chief product. A portion of the
escaping gas, on being tested, was found to contain a large proportion of
oxygen, thus proving indirectly the presence of nitrosyl chloride. The
product was a very volatile, mobile, crimson liquid, and in order to
ascertain its composition a portion was distilled through a series of
large thin glass bulbs, previously weighed. When about full, the
bulbs were sealed, temperature and pressure being noted, and weighed
again.

The contained vapour was now absorbed by water, a process
accomplished by connecting the two extremities a and b of the bulb
with burettes A and B, filled with water, the intervening tubing
being also quite full. The point a was then broken off inside the
tube, and the vapour was rapidly absorbed. When absorption was
complete the point b was broken, and a little more water sucked over
to wash out the tube a.

Burette A was now completely filled and inverted in a beaker of
water, and the levels being adjusted, a mark was made at c where the
water stood in the narrow tube. Burette B was now read, and from it
water was run into the bulb (through b) till its level rose to the mark ¢,
when B was again read. This gave the volume of unabsorbed gas, the
greater part of which was now contained in the burette A. The contents
of the bulb were then washed into a beaker for the determination of the
chlorine. The volume of gas in A was read and the nitric oxide then
absorbed by ferrous sulphate solution, after which the volume of

228 WILLIAMS: REACTIUNS SUPPOSED TO YIELD

residual gas (nitrogen) was again read. From these readings, it was
known how much nitrogen and, consequently, how much air had been
unexpelled from the bulb during distillation; this volume was sub-
tracted from the total capacity of the bulb, which was now measured
by running in water from a burette.

The data thus obtained furnished material for the determination of
the vapour-density and of the chlorine present in the gas. The
following table contains the numbers. Of the bulbs, No. I is nearest
the distilling flask, and contains therefore the least volatile portion of
the vapours.

Sertes A.—Bulbs sealed at 18°5° and 728°8 mm.

1°4

117 °75 174°6

18 °1513 23 *8571
17 *9341 23 -5376

0 +2172 0°3195

0°1367 0°20275
0 °3539 0°52225
0 *3076 0 °4794

37°35 37°17

Cl per cent. +| 21°5 | 22°71

Vo = vol. of bulb; Va = vol. of unabsorbed air; Vy = vol. vapour;
By = weight of bulb full of vapour; Ba = weight of bulb full of air;
Wr a = weight of vapour — weight of equal vol. of air;

Wa = weight of equal vol. of air (cale.) ; W» = weight of vapour.

In order to interpret these results, the following assumptions and
calculations were made.

Hypothesis «—If we assume that all the chlorine is combined as
NOCI, the volume of the latter can be calculated from the percentage
of chlorine, and hence the partial pressure on the admixed peroxide.
Then by means of the formula and tables of Professor J. W. Gibbs
(Chem. News, 1879, 298) the extent of the dissociation of the peroxide
can be determined. From this, the percentage composition of the gas
in the bulb is known, and the density of this mixture is then readily
obtained by a very simple calculation.

NITROXYL OR NITRYL CHLORIDE. 229

Hypothesis 8.—The chlorine is supposed to be free, and the calcula-
tion made in the same way.
Hypothesis y.—Nitroxyl chloride is supposed to be the chlorinated

substance.

The calculated and observed numbers are contained in the following
table, but the figures for the density of the gas in the first bulb
obtained on hypotheses 8 and y, were so wide of the mark that it was
thought unnecessary to work out the results for bulbs II and
III :—

Series A.

Il. III.

| Densities. iff. Densities. iff. Densities.

37°35 36°95 | — 37°17

37 “88 ' 38°08 ' 37 “92

40 92 ' = --
00 “6: - ~

There can be little doubt therefore that when chlorine and nitric

peroxide are passed through a red-hot tube the product is a mixture
of nitric peroxide with nitrosyl chloride, although no doubt some

chlorine is present in both in solution.

B. In these experiments, the mixed gases, obtained as before, were
led through a large U-tube filled with broken glass and heated in an
air-bath to a temperature varying from 130° to 150°. This tempe-
rature was chosen as being about that at which nitric peroxide
completely dissociates into 2NO,, and at which therefore one would
naturally suppose that direct combination with chlorine would most
readily take place.

The product of the experiment was a red volatile liquid similar to
that obtained in series A, but since no oxygen could be detected in the
escaping gas, we must conclude that dissociation of NO, into NO and
O; had not taken place, and, by implication, that nitrosyl chloride was
not formed in these experiments.

Vapour-density determinations were performed exactly as before,
except that the distillation of the vapour through the bulbs was con-
tinued for a very considerable time, the object being to ascertain, by
the action of the escaping vapour on glacial acetic acid, whether free
chlorine were present.

I do not consider this series of determinations to be as accurate as

VOL. XLIX. R

230 WILLIAMS: REACTIONS SUPPOSED TO YIELD

the next (C), as, owing to the purpose just alluded to, the distillation
was pushed so far that a considerable quantity of liquid condensed in
the first bulb, which was only removed with great difficulty.

The glacial acetic acid through which the issuing vapours were
bubbled was contained in a small distilling flask. After the operation,
the acid had a deep crimson colour, which, however, completely dis-
appeared when dry air was led through the acid for a short time at a
temperature of 100°. A portion of the acid was then diluted, and
nitric acid and silver nitrate added, the precipitate (a very small one)
being filtered off. The liquid was subsequently treated with excess of
yotash, and after boiling for some time the solution was acidified with
mitric acid. A considerable quantity of silver chloride remained un-
dissolved.

It is therefore certain that a notable amount of chloracetic acid had
been formed, even in ordinary daylight and at ordinary temperatures,
from which fact we must draw the inference that free chlorine was
almost certainly present in the vapeurs.

The following are the numbers for this series :—

Series B.—Bulbs sealed-at 24° and 761°7 mm.

III.

175 °O (approx.)
0°0

eeeeeeee

Bo ce cccces 73 °7767 60 *8276 79 *6245
Ba ....+e++| 73°5345 60°5154 79° 2724

Wrea ..-e. 0° 2422 0°3122 0°3521
Wa seccees 0°14043 0°184775 | 0° 20845

Wo ccccee’ 0°38263 0 -496975 0°56055
38°80

38 *82

39°33

5°66 5°5 —

During the manipulation of bulb III, it was broken, and conse-
quently no chlorine determination could be made. Its volume was
taken as being approximately 1 c.c. less than in series A, in order to
obtain an approximate vapour-density, as it was thought that possibly
this might be useful.

Calculations were made as in series A, the nitrosyl chloride

hypothesis being now excluded.

NITROXYL OR NITRYL CHLORIDE.

Serizs B.

Densities. Densities.

38°82
41 “64
41°73

DR. 6060200040460
Cale. hyp. 8

me 99 Yorcerceccos

There is here a slight difference in favour of 8, but considering how
small it is and how widely both series of calculated numbers vary
from the observed density, the figures cannot be taken as proving
anything.

C. The liquid for use in this series was prepared in the same way
as for the last, but the distillation was allowed to go on for a short
time only, in order that the most volatile portions of the vapour
should have influence. Hence the large amount of unexpelled air and

high percentage of chlorine.

Series C.—Bulbs sealed at 20°1° and 748°3 mm.

III.

144

..| 117°9 z
“55 . 18 *85

Cl per cent.. |

117°35

74 5864
74° 3507
0 °2357
0°13926

0°37496
0°1527

38°87
25°12

0°17824

0°47554
0 °2206

88°51
28°68

125°15

65°5103
65 *2689

0°2414
0° 14852

0 °38992
0°2219

37°9
35°11

The numbers in this series are, I think, fairly trustworthy, as

everything worked very smoothly throughout.
The following are the calculated results :—

WILLIAMS: NITROXYL OR NITRYL CHLORIDE,

Here there is a decided difference in favour of hypothesis 8. The
extreme variation in the differences is only 0°11, though it must be
admitted that the differences themselves are rather large.

Now in the calculated values on hypothesis +, the variation in the
differences amounts to 0°59, and the whole series is from 0°48 to 0°96
farther from the truth than the other. These discrepancies are con-
siderable, occurring as they do in a total difference never exceeding
2°6.
With regard to the uniform difference of 1°5 between the observed
densities and the numbers obtained on hypothesis 8, I think it is
easily understood when we remember that Gibbs’ formula is only
approximate, and that a like variation, the same in direction and
nearly the same in amount, was observed by Ramsay and Cundall in
calculating the density of a similar mixture with the help of this
formula.

From these experiments, I think we may reasonably conclude that
direct combination of nitric peroxide and chlorine, like all the other
methods, fails to yield nitry] chloride; the product hitherto regarded
as such being merely a solution of chlorine in nitric peroxide, together
with nitrosyl chloride, when the experiment is performed at a red
heat.

In two samples of this product, prepared at 150° on different occa-
sions, the chlorine was determined, and was found to be 8°6 and 9°2
per cent. respectively.

Now if a compound were formed, we should expect the amount of
chlorine to vary in products formed at different rates and during
different times. We must, therefore, I think, regard these numbers
as expressing the solubility of chlorine in the peroxide at the tempe-
rature of the condenser, which varied probably from — 20° to — 15°.

Indeed, a solution containing 4°75 per cent. of chlorine was readily
prepared by cooling the test-tube containing the peroxide during the
passage of the chlorine, and though the solution was probably not
saturated, nor was it so cold as in the spiral condenser, still the

MASSON ON SULPHINE SALTS. 233

liquid after this treatment strongly resembled that which dripped
from the delivery tube of the condenser.

I must mention, however, that when endeavouring to prepare a
specimen of this solution a few days ago for exhibition to the Society,
the peroxide, though it absorbed the chlorine with great vigour,
underwent no appreciable change of colour, nor was any such change
induced when a portion of the liquid was warmed slightly. The
temperature of the experiment was — 15°, and the peroxide was
quite free from water, as was shown by its solidifying during the
passage of the chlorine.

I am inclined to think that a trace of moisture is necessary for the
development of the red colour, which is probably due to a small
quantity of nitrosyl chloride :—

Cl, + H,O = 2HCl + 0; HCl + (NO)NO; = NOC! + HNO,.

In conclusion, I must render my hearty thanks to Dr. Tilden
especially, and also to Dr. Nicol, for their ready and kindly help.

XXVI.—On Sulphine Salts containing the Ethylene Radical. Part I.
Diethylenesulphide-methyl-sulphine Salts.

By Orme Masson, M.A., D.Sc., University of Edinburgh.

Tue discovery made by von Oefele in 1863 (this Journ., 17, 106),
that ethyl iodide and ethyl sulphide combine directly to form the
iodide of a sulphine base, led to the investigation of the products of
several similar reactions and to the establishment of many points of
analogy between the sulphine salts and the salts of organic ammonium
and phosphonium bases. Still, the chemistry of sulphines is as yet
very incomplete. Nothing, for instance, is known of di- and tri-
sulphines comparable to the di- and tri-ammoniums; and very little
of monosulphines containing other than monad radicles. Attempts
were, indeed, made by Dehn and by Cahours to prepare sulphines
containing the ethylene radical very soon after the publication of von
Oefele’s research; but these attempts were not successful, though
they led to other interesting results. Their work will be more specially
dealt with in Part II of this paper; here it may suffice to state that
the method which they both employed — the heating of ethylene
bromide with an alkyl sulphide—was proved by Dehn to result in

234 MASSON ON SULPHINE SALTS

double decomposition between the reagents, followed by combination
of the alkyl sulphide with the alkyl bromide so formed :—

2C.H,Br, oa 2(CH;).8 = (C.H,).8, + 4CH,Br.
CH,Br + (CH;).8 = (CH;);SBr.

It occurred to me that ethylene sulphine salts might be obtained
by acting on ethylene sulphide with the haloid compound of methy]
or ethyl, as in that case probably there would be no such tendency to
double decomposition.

Constitution of Diethylene Disulphide.

There are in reality two distinct substances which have the
empirical formula C,H,S. One of them, discovered by Léwig and
Weidmann in 1840, and afterwards investigated by Crafts (Annalen,
128, 220), isa white amorphous powder produced by the action of
potassium sulphide on ethylene bromide. It is insoluble in all
ordinary solvents, and cannot be volatilised without undergoing
chemical change. Nothing is known of its true constitution. Very
different from this is the diethylene disulphide discovered in 1863
quite independently by Crafts (Annalen, 124, 120; 125, 123) and
Husemann (Annalen, 126, 269). This is a white crystalline substance,
soluble in alcohol, ether, benzene, and carbon disulphide, volatile even
at ordinary temperatures, melting at 111°, and boiling at 200°. It is
prepared from the amorphous sulphide by heating it at 160°, either
alone or with carbon disulphide in sealed tubes; and I have found
that it is also directly formed in considerable quantity, together with
the amorphous sulphide, when strong alcoholic solutions of ethylene
bromide and potassium sulphide are mixed. The following is a
summary of what is known regarding its constitution.

Crafts, who regarded it as C,H,S, showed that it is a true sulphide,
analogous to those of the alcohol radicals, by preparing its halogen
and oxygen addition products. Thus he obtained the bromide
C,H,SBr, by direct addition of bromine, the sulphoxide C,H,SO by
decomposing the bromide with water and by the action of nitric
acid on the sulphide itself, and the sulphone C,H,SO, by the further
action of nitric acid. Husemann, however, ascertained that the
vapour-density of the sulphide requires its formula to be doubled ;
and hence he represented it and its derivatives as (C,H,).S.,
(C.H,).8.Br,, (C2H,4)28.02, (C2Hy)2S,0,. If these formule are correct,
it is evident that the molecule of the sulphide must be symmetrical,
and must contain two dyad sulphur-atoms, conditions that are well
fulfilled by the formula given in some of the text-books,

OH,
S<on®-

CONTAINING THE ETHYLENE RADICAL. 235

Dehn’s theory (Annalen, Suppl. 4, 83) that it is the sulphide of a
sulphinic base, (C,H,).S:S", must be at once rejected; and the
formula C.H,: S : S : C.H,, suggested in Watts’ Dictionary (6,
p. 607), is equally incorrect, as it does not allow of the formation of
a disulphone. Moreover, the constitution given above agrees well
with two reactions which yield the sulphide, discovered by Husemann;
viz., that between ethylene bromide and the mercury:salt of ethylene
mercaptan, and that between ethylene bromide and: ethylene thio-
carbonate :—

CH.<S>Hg + BrOsH, = 0.H,<$>C,H, +-HgBr, and

C.H.<g>CS8 + BrC.H, = GH.<g>C.H. + OSBr,

(the thiocarbonyl bromide being represented by the products of its
decomposition).

It is evident, however, that these reactions are quite compatible
with the view that the sulphide is, as Crafts represented it, C,H,S;
and that, since none of its derivatives require the double formula, the
proof that it is (C,H,).S., and not C,H,S, has hitherto rested solely on
the physical evidence of its vapour-density. Though this is generally
accepted as sufficiently convincing, it is satisfactory to obtain cor-
roborative chemical evidence such as is given in the following pages,
where derivatives are described which do not admit of the. halving of
the formula (C,H,).S83.

Preparation of Diethylene Disulphide.

The following method has been found to work easily for tho
preparation of this substance in quantity. Ethylene bromide, dissolved
in about its own volume of alcohol, is mixed in a large flask with a
concentrated solution of potassium sulphide, more than.an equivalent
quantity of the latter being employed. The flask is connected with
along upright condenser, and should be placed in a draught cupboard,
as the odour evolved (due to the formation of the mercaptan and
other bye-products) is extremely unpleasant. The action begins at
once, the amorphous ethylene sulphide and potassium bromide being
precipitated ; and so much heat is developed that ebullition occurs in
a very few minutes. When the reaction is over and the flask has
become quite cold, enough water is added to give fluidity to its almost
solid contents, the condenser is placed in the normal position, and a
current of steam is led in by a tube dipping nearly to the bottom of
the flask, which is at the same time heated on a sand-bath. The
distillate which is collected cousists at first of dilute alcohol con-

236 MASSON ON SULPHINE SALTS

taining diethylene disulphide in solution. As the alcohol becomes
weaker, the disulphide is seen to solidify in the condenser in the form
of crystals, which are washed down into the receiver. The distillation
is continued till no more of the disulphide passes over; the alcoholic
distillate is then largely diluted with water; and the precipitated
disulphide is collected and dried by pressing between folds of blotting-
paper, and, finally, over sulphuric acid. The contents of the flask,
which still contains the amorphous sulphide, are poured on to a filter ;
this is well washed with hot water, and dried at 100°.

The fact that the crystalline disulphide is produced by the reaction
between ethylene bromide and potassium sulphide seems to have been
overlooked by previous observers, who also appear to have been
unaware of its convenient property of volatilising in the vapour of
water or alcohol at a temperature much below its own boiling point
(200°). That it is really produced simultaneously with the amorphous
sulphide, and not from it by the action of the steam, is proved by the
limited amount obtained, and by the fact that this change from the
amorphous to the crystalline state does not occur below about 160°,
either in a sealed tube containing water, or under any other known
conditions.

The next step in the preparation is to induce this change in the
amorphous product. Husemann recommended heating it in sealed
tubes at 160° with carbon disulphide; but this is a troublesome
operation, and altogether unnecessary. A fairly good yield is obtained
by simply heating the amorphous powder in a large retort, the lower
part of which dips into an oil-bath, provided that the temperature is
raised very gradually and is then kept several hours at 160°, but not
allowed to rise much above that temperature. The product sublimes
into the upper part and neck of the retort, where it slowly forms
a hard yellow glassy cake. A certain amount of charred matter
remains at the bottom of the retort when the operation is over. The
cake of crude disulphide is then shaken out, broken up, washed
sparingly with alcohol, which frees it from most of its objectionable
odour (due to an oily bye-product), and is distilled in a current of
steam. After pressing and drying, it is quite white and pure, but
the water remaining in the distilling flask contains much oily and
disagreeably smelling impurity.

In one preparation, I used 650 grams of potash (converted into
potassium sulphide) and 1000 grams of ethylene bromide; and
obtained, in the first stage, 45 grams of the crystalline disulphide
and 126 grams of the amorphous sulphide. This is 54 per cent. of
the theoretical yield of C,H,S from C,H,Br, (319 from 1000); but
better results were got on a smaller scale by the use of a larger
excess of potassium sulphide. In the second stage, the 126 grams of

CONTAINING THE ETHYLENE RADICAL. 237

the amorphous compound gave 75 grams of pure diethylene disulphide;
so that, in all, 120 grams of this substance were prepared from 1000 of
ethylene bromide, that is, 37 per cent. of the theoretical yield.

Action of Methyl Iodide on Diethylene Disulphide.

It was not unnaturally expected that, if any combination of these
two substances occurred, 1 mol. of the disulphide would combine with
2 mols. of methyl iodide, and that a sulphine salt would be formed of
the constitution—

H;C C,H CH;
"T>8<oHS<r *
This, however, does not take place; for only 1 mol. is taken up, even
when a large excess of the iodide is employed. The salt produced is
the iodide of a base which may be viewed as trimethylsulphine in
which two of the methyl-groups are displaced by the dyad radical,
S(C,H,;)"2,—a view that is best expressed by the name diethylene-
sulphide-methyl-sulphine :—

C,H, CH;

H.C. . CH
H.c> ><], * S<oH oS

Trimethylsulphine Diethylenesulphide-methyl-
iodide. sulphine iodide.

Such a base should resemble trimethylsulphine in its general
characters, but should differ from it in certain respects on account of
the comparatively negative character of the S(C,H,), group; which
ought to impress itself (so to speak) on the whole molecule. This, as
will be shown, is actually the case.

The reaction occurs slowly at the ordinary temperature. When
ethereal solutions of the two substances are mixed in a flask and set
aside for some weeks, minute crystals of the sulphine iodide gradually
separate and form a crust on the sides of the flask. Ata tempera-
ture of about 70°, however, the reaction is completed more rapidly.
The mixture of diethylene disulphide and methyl iodide (the latter
in slight excess of the theoretical quantity) is heated for about 12
hours at that temperature in sealed tubes placed in a water-bath.
Some pressure is developed, and a combustible gas escapes when the
tubes are opened, after cooling. The crude product of the reaction is
a solid reddish-coloured mass, and consists of a mixture of two
substances,—the iodide and the tri-iodide of diethylenesulphide-
methyl-sulphine. After the excess of methyl iodide has been got rid
of by powdering in a mortar and exposure to the air, these two salts
may be separated by washing with boiling alcohol, which dissolves
out the dark-red triiodide and leaves the white iodide unaffected ; or

238 MASSON ON SULPHINE SALTS

the whole ma3zs may he treated with a somewhat large volume of
boiling water, which dissolves the iodide and throws down the triiode
as a black tarry precipitate. The solution is filtered, and is then
boiled in a beaker till it has lost its yellow colour and been reduced
to a comparatively small bulk; but the evaporation must not be
carried so far as to cause decomposition. On cooling, it deposits a
plentiful crop of opaque white apparently cubical crystals; and
successive crops of these may be obtained by alternate evaporation
and cooling of the mother-liquor. The total yield of iodide obtained
in this way is nearly equal to double the weight of the diethylene

disulphide employed.

Diethylenesulphide-methyl-sulphine Salts.

The iodide, C;H,S,I. After being purified by recrystallisation
from hot water, the salt was analysed with the following results :—

I, 0°2087 gram gave 0°1745 gram of CO, and 0°0818 gram of H,0.
IT. 0°1981 ~ 0°1662 - 0°0775 7
IIT. 0°1793 gram gave 0°3142 gram of BaSQ,, after ignition with a
mixture of sodium carbonate and potassium chlorate.

IV. 0°1993 gram gave-0°1783 gram of AgI.

Found.
Calculated for -—— ~ —_—T
C;H,,8,L. I. II. III. IV.
svssec 22°90 22-80 22°88 —_ —_
DE étaeen 4°20 4°30 4°35 _ —

The salt is readily soluble in hot water, much less so in cold water,
very sparingly in alcohol, and insoluble in ether. From hot water, it
crystallises in opaque white apparently cubical crystals; but the cold
aqueous solution evaporated in a vacuum deposits it in the form of
transparent needles. The aqueous solution undergoes some decom-
tion when evaporated to dryness at 100°. The dry salt sublimes
without melting when heated somewhat above 100°; and is at the same
time completely decomposed, partly, at any rate, into its original
constituents, for crystals of diethylene disulphide are formed in the
upper part of the vessel in which the operation is conducted. In this
decomposition, the behaviour of the salt is analogous to that of
trimethylsulphine iodide. Towards silver salts, it also conducts itself
in a similar manner, giving silver iodide and a new salt of the
sulphine base. In order, however, to obtain the hydroxide by the

CONTAINING THE ETHYLENE RADICAL. 239

action of moist silver oxide, it is necessary to operate on a very dilute
cold solution of the iodide; for the reaction will not take place in a
strong cold solution, and any application of heat at once decomposes
the base with the formation of silver sulphide and a disagreeably
smelling volatile liquid. It is therefore more convenient to prepare
the other salts by acting on the iodide with the corresponding salt of
silver or cn the sulphate (so obtained) with the corresponding barium

salt.
The triiodide, CsH,,S.1;, which, as already stated, occurs as a bye-

product in the formation of the iodide, is easily obtained by the direct
combination of the latter salt with iodine. For this purpose, the
calculated quantities are weighed out, the iodide is dissolved in
sufficient hot water, and the iodine in sufficient hot alcohol, and the
two are mixed. When the solution cools, it deposits the triiodide in
the form of very thin plates of a lustrous garnet-red colour. They
may be purified by-washing with ether and recrystallising from hot
alcohol. The following are the results of analysis :—

I. 0°4255 gram gave 0°1821 gram of CO, and 0°0835 gram H,0.
II. 0°4364 gram, after treatment with sulphurous acid solution,

gave 0°5958 gram of AglI.

Calculated for
C,H))821.

The salt is insoluble in ether, only slightly soluble in cold alcohol,
freely in hot alcohol. When warmed with water, it melts and
aggregates into heavy black oily drops; these gradually decompose, on
boiling, into the original iodide and iodine. Sulphurous acid solution
acts on it as on a mixture of these substances. When heated by itself
it melts at about 89° ; and at about 145° it begins to evolve a brown
oily distillate, whilst at a still higher temperature a white crystalline
sublimate and iodine vapour are formed. Ammonia solution at once
converts it into a black amorphous substance which explodes when
heated, but not so violently as does iodide of nitrogen, to which it
bears a close resemblance.

The formation of this triiodide in the preparation of the iodide is
partly, no doubt, accounted for by the presence of free iodine in the
methyl iodide employed; but to a larger extent it must be due to a

240 MASSON ON SULPHINE SALTS

secondary action, which gives rise also to the combustible gas whose
presence is always noticed on opening the tubes.

Two distinct views are possible as to the constitution of this
substance. The first would represent it as formed by the union of
iodine with the dyad sulphur-atom; in which case it would be
analogous to the tetrahaloid derivatives of diethylene disulphide,

thus :—
C.H, C.H, CH,
LS<¢iy'>Sh LS<Gig'>S<|
Diethylene-disulphide tetr- Diethylene-disulphide-methyl-
iodide. sulphine tri-iodide.

As is well known, however, various organic bases in whose molecules
there are no atoms comparable to this second sulphur-atom, form
polyiodides; and it was shown by Dr. Dobbin and myself (Trans.,
1885, 56) that trimethylsulphine iodide combines directly with
iodine, bromine, and chlorine to form definite crystalline compounds of
the general formula (CH;);SX;. Now the salt under consideration so
exactly resembles these compounds in its properties (more particularly
in its behaviour with water and with ammonia), that it seems right to
conclude that its constitution is similar to theirs (as yet not deter-
mined with certainty), and that it should be represented in the mean
time as S(C,H,).S(CH;)I,L, not as I,8(C.H,),.SCH,I.

In this connection, it appeared to be a matter of some interest to
ascertain whether the iodide would combine in a similar manner with
bremine and chlorine. According to the view of its constitution
taken in this paper, viz., that it is trimethylsulphine iodide in which
two methyl-groups are displaced by the comparatively negative
group S(C,H,)’, one would expect it to show less tendency than
trimethylsulphine iodide itself to combine with strongly negative
radicals such as the halogens. This is confirmed by experiment.
Dry bromine vapour passed over the salt is absorbed, the white
iodide assuming a scarlet colour, but without melting (in the case of
trimethylsulphine iodide much heat is developed and the salt melts) ;
but this scarlet product is not very stable, for it decomposes in the air,
and rapidly when heated with alcohol or water, in which last case a
solution of the sulphine bromide is obtained. Chlorine, on the other
hand, has absolutely no action on the iodide if both are dry; whilst
it combines very energetically with dry trimethylsulphine iodide.

The nitrate, C;sH,,S,NOs, is obtained by mixing a solution of silver
nitrate with that of the iodide. The point of exact equivalency can
be very nearly hit by observing the colour of the silver precipitate,
which changes from a yellowish-white to a distinct yellow as soon as
sufficient silver nitrate has been added. The aqueous solution under-

CONTAINING THE ETHYLENE RADICAL. 241

goes some decomposition when evaporated to dryness; but the pro-
duct is easily purified by crystallisation from hot alcohol.

0°2924 gram gave on combustion 0°3232 gram CO, and 0°1561

Calculated for
C;H,,S.N03. Found.

30°15
5°93

The nitrate is very soluble in water, but is not deliquescent. It is
much less soluble in cold alcohol than in hot, and is insoluble in ether.
It separates in the form of thin pearly plates when the hot alcoholic
solution is cooled or when ether is added to the cold alcoholic solu-
tion; whilst by slow evaporation it is obtained in large apparently
rhombic crystals. It melts at 172°, and at once begins to decompose,
a crystalline sublimate being formed.

The silver salt, CsH1,S.NO;,AgNOs;, is obtained by mixing a strong
solution of the nitrate with an excess of silver nitrate and then pre-
cipitating with alcohol. It forms small colourless barb-like crystals.
It blackens on exposure to sunlight, and detonates slightly when
heated. It is extremely soluble in water, less so in alcohol, and in-
soluble in ether.

0'1517 gram on ignition left a residue of 0°0449 gram of Ag.
Calculated for
C;H,,S.NO3,AgNOs. Found.
29°60

The sulphate, 2(C;HyS2)280.,7H,0. The salt is obtained by treat-
ing the iodide with silver sulphate solution, evaporating the filtered
liquid to a syrup, dissolving this in strong alcohol, and either preci-
pitating with ether or crystallising out ina vacuum. It is a remark-
able fact that throughout this treatment the salt retains its water of
crystallisation; indeed it does not appear capable of parting with
it in any circumstance without undergoing further change. The
decomposition produced by evaporation of the aqueous solution at
100° is greater in the case of the sulphate than in those of the other
salts; it results in the formation of a flocculent white insoluble
substance, free sulphuric acid, and a volatile sulphur compound
having a very nauseous odour. There will be occasion again to
mention this odour and the substance which gives rise to it. The
sulphate crystallises in large deliquescent prisms when the aqueous
solution is slowly evaporated over sulphuric acid in a vacuum, in
small needles when the alcoholic solution is similarly treated, and in
large feathery groups when ether is allowed to diffuse down into the
cold alcoholic solution. When heated, it melts with decomposition at

242 MASSON ON SULPHINE SALTS

127°, water being given off together with diethylene disulphide and
other volatile matters; and at a higher temperature charring occurs.
I. 0°2470 gram gave 0°2512 gram CO, and 0°1471 gram H,0.
If. 0°2457 - 0°2476 - 0°1451 “
IIT. 0°3644 - 0°3693 i 0°2149 a
IV. 0°2508 gram gave 0°1381 gram BaSO, by precipitation with

bariam chloride.
V. 0°2425 gram, treated in the same manner, gave 0°1324 gram

BaSQ,.
Found.
Calculated for r a -
2(C,H,,8,).80,,7H,0. | a II. III. IV. V.
ee 27°97 27°73 27°48 27°64 —
6°76 662 656 6:55 —_ _

Different preparations were used in this analysis, but both were
from alcoholic solution. III and V were with one preparation ; I, II,
and IV with another. The hydrogen is obviously too low; but the
formula given agrees much better than any other with the figures

obtained.
The chloride, C,H,,S,Cl, may be prepared from the sulphate by

exact precipitation with barium chloride and evaporation of the
filtered solution. It is readily soluble in water, but slightly so in
alcohol, and insoluble in ether. The hot saturated aqueous solution
deposits a magma of needles on cooling ; whilst the salt is obtained in
the form of large transparent tables by the slow evaporation of the
cold aqueous solution. It is not deliquescent.
0°2393 gram gave 0:2011 gram AgCl.
Calculated for

C;H,,S,Cl. Found.
Cl ..sccreeee 20°82 20°86 per cent.

Further analysis was not considered necessary, more especially as
the platinichloride* has been fully analysed.

The chloride forms several characteristic compounds with metallic
salts. Of these the platinum, gold, and mercury compounds have

been examined.
Platinum salts.— Diethylenesulphide-methy]-sulphine chloride reacts

* The term “ platinichloride,” as being more systematic, is here applied to the
compounds containing platinic chloride, PtCl,.

CONTAINING THE ETHYLENE RADICAL. 243

with platinic chloride in a peculiar and interesting manner. As
many as four distinct salts have been prepared and analysed. Of
these four, three are direct compounds of C;H,,8,Cl with PtCl, in
different proportions, whilst the fourth is a similar compound minus
the elements of hydrochloric acid. The salts are—

Salt. Chlorine ratio.
2C;H,,8,Cl,PtCl, l:
4C,H,,8,C1,3PtCl, 1:é
C;H,,8.Cl, PtCl, 1:
C,H,,S,Pt,Cly me

It is obvious that the first three, as judged by the ratio of the
chlorine in the one half of the molecule to that in the other, form a
definite series, whilst the fourth is quite distinct in character. If
judged by their properties, however, the salts must be differently
grouped ; for the second, third, and fourth are insoluble amorphous
yellow powders, whilst the first is orange-coloured, crystalline, and
soluble.

The platinichloride, 2C;H,,S,Cl1,PtCl, is obtained as a heavy
orange crystalline powder when a cold strong solution of the sulphine
chloride is precipitated with platinic chloride. The salt is insoluble
in alcohol and ether; sparingly soluble in cold water. It cannot
be dissolved in hot water, as this converts it into the insoluble
salt C,H.,S,Pt,Cl,; but it may be dissolved unchanged, in part
at any rate, in a hot aqueous solution of the sulphine chloride;
and from such a solution it separates on cooling in the form
of groups of interlacing golden-yellow needles. The cold aqueous
solution deposits, on standing, an insoluble amorphous yellow powder,
which was found to contain, in one instance, 36°35 per cent. of
platinum, and is most probably a mixture of the salts C;H,,8,Cl, PtCl,
(which requires 38°42 per cent.) and 4C,;H,,8,C1,3PtCl (which
requires 34°55 per cemt.). Analysis of the platinochloride gave the
following results :—

I. 0°4264 gram gave on combustion 0°2714 gram CO, and
01277 gram H,0, and left a residue of 0°1232 gram Pt.

II. 02512 gram gave 0°0720 gram Pt.

III. 0°4537 - 0°1303 -

IV. 02535 gram, ignited with sodium carbonate, gave 03188
gram AgCl.

VY. 0°2991 gram, ignited with sodium carbonate and potassium
chlorate, gave 0°4076 gram BaSQ,.

MASSON ON SULPHINE SALTS

Found.
Calculated for _ SS
2C,H),S,Cl,PtCl,. ) 3 III.
17°36 —
3°33 —-
28°89 28°67 28°72
— — — 31°11

Before being analysed, the crystalline precipitate was washed on a
filter with cold water, alcohol, and ether, and dried at the ordinary
temperature in a vacuum. When it has been freed from moisture in
this way, it may safely be heated at 100°; above 130°, it turns black
and decomposes without melting.

The salt CiH»S,,Pt,Cl, is produced, as stated above, when the
platinichloride is digested with boiling water. The orange salt at
once loses its crystalline character and assumes the appearance of a
yellow amorphous insoluble powder. The aqueous filtrate from this
has a distinctly acid reaction, due to free hydrochloric acid. Evapo-
rated to dryness on the water-bath, it yields a crystalline residue of

diethylenesulphide-methyl-sulphine chloride, which is readily identi-
fied by its reactions with the chlorides of gold, platinum, and mercury,
by boiling it with soda solution (see later). The reaction which
occurs may, therefore, be expressed by the equation—

2(2C;H,,S,Cl,PtCl,) = CyHaS,Pt.Cl, + HCl + 2C;H,S,Cl.

The analysis of the insoluble salt gave the following results :—

I, 0°2874 gram gave 0°1287 gram CO, and 0°0611 gram H,0.
II. 0°2656 - 01050 ,, Pt.
ITT. 0°2271 " 00898 , ,,
IV. 0°2510 gram, ignited with lime, gave 0°3294 gram AgCl.

Found.

Caleulated for
C)pHS8,Pt2Cly.

32°65
100-00
The two salts, 4C,H,,8,C1,3PtCl, and C;H,,8.Cl,PtCh, both of

* The atomic weight of platinum is taken as 195.

CONTAINING THE ETHYLENE RADICAL. 245

which are insoluble amorphous yellow powders, closely resembling
the salt C,)H,,S,Pt,Cl, in appearance and properties, are thrown down
as a mixed precipitate when platinic chloride is added to a hot solution
of the sulphine chloride. Such a precipitate was found to contain
37°61 and 37°70 per cent. of platinum. Digesting it with boiling
water and with alcohol did not affect its composition. As the formula
CsHy,8,Cl,PtCl, requires 38°42 per cent. of platinum, this precipitate
was suspected of containing it tegether with some impurity; an
attempt was therefore made to obtain it in a pure condition by frac-
tional precipitation. Accordingly, a weighed quantity of the pure
sulphine chloride was dissolved in boiling water, and hot platinic
chloride solution—an amount very nearly equal to that calculated for
the formation of the salt C;H,,S,Cl,PtCl,—was added in four successive
equal portions. The first two fractions, taken together, gave
34:14 per cent. of platinum; the third gave 3406; and the fourth,
3426. The orange-coloured filtrate from this last fraction was again
boiled, when it almost at once threw down a fifth precipitate, larger
than all the rest put together, and of a slightly different shade of
yellow. This, on analysis, proved to be the salt C;H,,S,Cl,PtCl,, in a
fairly pure state.

I. 0°2976 gram gave 0°1242 gram CO,, 0°0632 gram H,0, and

01141 gram Pt.
II. 0°1957 gram gave 0°0751 gram Pt.
IIl. 0'2151 gram, ignited with lime, gave 0°2966 gram AgCl.

Found.

Calculated for
C;H,,8,Cl, PtCl,.

34°11

The smaller fractions, after the separate platinum determinations
had shown them to have practically the same composition, were
mixed for further analysis.

I. 02558 gram gave 0°1364 gram CO, and 0°0638 gram H,0 (the
platinum was lost).
IT. 01946 gram, ignited with lime, gave 0°2620 gram AgCl.

III, IV, and V (separate fractions), 0°1245, 0°1324, and 0°1608 gram

respectively gave 0°0425, 0°0451, and 0°0551 gram Pt.

VOL. XLIX.

MASSON ON SULPHINE SALTS
Found.
III. TV.

Calculated for
40,H,,8.Cl,,3PtCl,. f.
14°54

2°77
_ — 3414 3406 3426
33°55 — 3330 — — —

100-00

The substance was thus proved to be the salt 4C;H,,S,Cl1,3PtCl,
in a nearly pure condition.

These amorphous salts are practically insoluble in alcohol, ether,
water (hot or cold), and dilute acids. They are slightly soluble in
hot strong hydrochloric acid, and in strong aqueous ammonia, and
still more so in hot strong nitric acid; but they cannot be obtained
from these solutions in the crystalline state. Cold solution of hydric
sulphide does not affect them; but if they are suspended in water at
100° and submitted to the action of a current of the gas, they are
completely decomposed, platinic sulphide being formed. Ammonium
sulphide causes a similar decomposition at the ordinary temperature.
They behave like the orange salt when heated, turning black, and
quietly evolving volatile matter without melting.

The aurichloride, C;H,,8S,Cl,AuCl;,, is a light yellow amorphous
powder, thrown down when gold chloride is added to a cold solution
of diethylenesulphide-methyl-sulphine chloride. When heated with
water, it dissolves, but is at once decomposed ; and there gradually
separate, not crystals of the salt, but minute yellow spangles of
metallic gold. These impart to the liquid a greenish tinge, when
looked at by transmitted light. This forms a very characteristic test
for the chloride.

0°1808 gram gave on ignition 0°0761 gram Au.

Calculated for

C;H);S,Cl, AuCl,. Found.
42°09

The mercury salt, C;H,,S.Cl,HgCh, is produced as a crystalline
white precipitate, when strong aqueous solutions of the two consti-
tueuts are mixed. It may be recrystallised from hot water, when it
separates in the form of needles mixed with thin plates.

0°6191 gram gave 0°3318 gram of HgS.

Calculated for
C;H,,8,Cl, HgCl,. Found.
46°20

CONTAINING THE ETHYLENE RADICAL. 247

The hydrowide, C;H,S.°OH, has not been isolated ; but a solution of
it is formed when a cold and very dilute solution of the iodide is
shaken with recently precipitated silver oxide. It has an alkaline
reaction, and acts like an alkali on the salts of the heavy metals. It
cannot be boiled without undergoing decomposition; the products are
an insoluble white flocculent matter and a volatile sulphur compound
with a peculiarly disagreeable odour, the same that has already been
mentioned as characterising the decomposition of the sulphate and
occurring when a solution of the iodide is warmed with silver oxide.

In order to investigate these decomposition products of the hydroxide,
a solution of a weighed quantity of the sulphate was mixed with a
slight excess of barium hydroxide, and filtered, without the applica-
tion of heat. The filtrate quickly developed the peculiar odour, which
gradually increased in intensity ; and after it had stood for some hours
drops of an oily liquid began to separate out, thus showing that the
hydroxide decomposes slowly, even at the ordinary temperature. The
liquid was then distilled. The decomposition became complete just
before the boiling point was reached, and it was marked by the appear-
ance of a sudden white turbidity in the liquid. The oil all came over
with the distillate; and the residual liquid in the flask contained a
small quantity of a bulky white flocculent substance.

For analysis, the oil was simply dried with calcium chloride and
filtered, as it was found that it could not be distilled alone at the
atmospheric pressure without decomposition. The results obtained
point to the formula C,,.H,S;.

I. 0°2664 gram gave 0°4257 gram CO, and 0°1806 gram H,0.
IT. 0°2460 - 0°3927 ” 0°1638 .
IIT. 0°2620 gram, after ignition with sodium carbonate and mereuric
oxide, gave 0°9352 gram BaSQ,.

Found.
Calculated for cr A.
C,H2,8;. ie IL.
43°58 43:54
7°53 7°40

It is a mobile, highly refractive, colourless, very disagreeably
smelling liquid, not miscible with water, but soluble in alcohol and
ether. Its sp. gr. is 1°044 at 16° (water at 15°5° = 1). Its reaction
is neutral. When heated to 150°, it decomposes, becoming dark in
colour, without ebullition. It may, however, be readily volatilised in
a current of steam. It is vivlently attacked by nitric acid, which

248 MASSON ON SULPHINE SALTS

oxidises it first to a substance of the nature of a sulphone, and then,
by further action, to sulphuric and oxalic acids. It is readily acted
on by chlorine-water and bromine-water, with the formation of yellow
or red resinous substances. It also combines with methyl iodide, the
product being a syrapy liquid, which cannot be induced to crystal-
lise.

The white amorphous decomposition-product of the hydroxide,
formed at the same time as the oil, was collected on a filter, washed
with dilute hydrochloric acid and with water, and dried. The bulky
appearance that it presented while suspended in water was most mis-
leading, for its total quantity proved to be very small—too small to
permit of an analysis being made. It melts when heated on platinum
foil, evolves a sulphide odour, and then barns with a blue flame,
leaving no residue. It is evidently, therefore, an organic sulphur
compound ; in appearance, it resembles the amorphous ethylene sul-
phide. Though it is produced in such small quantity, its formation
is a constant and apparently necessary accompaniment to that of the
oil, and a knowledge of its mature is required before an equation can
be given for the decomposition of the hydroxide.

Action of Alkalis on the Salts of Diethylenesulphide-methyl-sulphine.
—If a solution of the iodide, chloride, sulphate, or nitrate, be heated
with potash, phenomena occur of a precisely similar nature to those
just described as resulting from the action of heat on the hydroxide.
The clear solution becomes suddenly turbid at a temperature near the
boiling point ; and the oil, C,,H,S;, distils over with the steam. The
change also occurs slowly at the ordinary temperature. Baryta-water
acts on the soluble salts in the same manner, but less readily. From
this it must be inferred that alkalis are capable of turning out the
sulphine base from its salts, or, to speak more accurately, of entering
into double decomposition with those salts. This reaction not only
affords a most convenient and delicate test for the sulphine radical
(the smell of the oil -being unmistakeable), but marks a distinction
of some theoretical interest between it and the trimethylsulphine
radical; for it is well known that the salts of the latter are not
attacked by alkalis, in which respect, as in many others, they
resemble the tetramethyl-ammonium and tetramethyl-phosphonium
salts. But this difference between the two sulphines is just what
might be expected if the view of the constitution of the diethylene-
sulphide-methyl-sulphime salts that is expressed in the name given to
them be correct; for it involves the assumption that their base is less
positive than that of the trimethylsulphine salts, though in most
respects strictly analogous to it.

It has already been mentioned that ethylene sulphide occurs in two

CONTAINING THE ETHYLENE RADICAL. 249

distinct forms, the crystalline disulphide, (C,H,),S., and the amorphous
variety, whose constitution is unknown, and which may be therefore
written (C,H,),S,. The m may be 1, or may be more than 2; that is
to say, the change which occurs when the disulphide is produced from
it by the action of heat may be a polymerisation, or a breaking up of
the molecule. It occurred to me that a test between these two possi-
bilities might be afforded by the action of methyl iodide on the
amorphous sulphide. If it be C,H,S, and if combination occurs at
all, one would certainly expect a sulphine salt of the constitution

C,H, = S58 a to be formed. This salt and its derivatives would

probably possess very similar properties to the trimethylsulphine salts,
from which they would also ditfer very little in composition. If, on
the other hand, the amorphous sulphide be more complex than the
crystalline one, a totally different result might be expected. It has
been shown that (C,H,).8, can combine readily with CH,I, but not
with 2CH;I; and it is but reasonable to conclude that (C.H,),S,,
where n is greater than 2, would behave in an analogous manner, and
that some of the sulphur-atoms in the molecule would remain
unaffected by the reaction. Of course the action must take place, to
be of any value as a test, at a temperature below that at which the
change from one form of sulphide to the other is known to occur, 1.e.,
below about 160°.

A preliminary experiment has shown that action occurs readily at
60—70° in a sealed tube. A combustible gas and tarry prodacts are
formed ; but, in addition to these, a sulphine iodide is produced which
may be extracted by warm water, and freed from impurities by filtra-
tion with animal charcoal. The salt obtained by evaporation of the
solution does not at all resemble diethylenesulphide-methyl-sulphine
iodide in solubility or crystalline form, but does bear a close resemb-
lance to the iodide of trimethylsulphine. I hope to be able to
investigate the reaction fully.

XXVII—On Sulphine Salts containing the Ethylene Radical.
Part II. Dehn’s Reaction between Ethylene Bromide and Ethyl
Sulphide.

By Orme Masson, M.A., D.Sc.

Iv 1865 Dehn published (Annalen, Suppl. 4, 83) an account of the
action which occurs when ethylene bromide and ethyl sulphide are

heated together with water in sealed tubes. In the same year,
s 2

250 MASSON ON SULPHINE SALTS

Cahours (Annalen, 136, 151; and Compt. rend, 60, 1147) described
the action of methyl sulphide on ethylene bromide. They arrived,
however, at very different conclusions; for, whilst Cahours stated
that a disulphine salt of the constitution (CH;),(C,H,)S.Br, is pro-
duced, Dehn observed no such product of direct union. According to
him, the main reaction consists in double decomposition by which
ethyl bromide and diethylene disulphide are formed—

I. 2C.H,Br, + 2(C.H;).8 — (C,H,).S.* + 4C.H;Br,

and this is accompanied by the union of part of the ethyl bromide
with ethyl sulphide to form triethylsulphine bromide—

II. C,H,Br + (C,H;).S = (C,H,),SBr.

Returning to the subject in 1869 (Ber., 2, 479), Dehn stated that
he had repeated his experiments with similar results, using methyl
sulphide in place of ethyl sulphide; and he accordingly asserted that
Cahours’ supposed disulphine salt was in reality only trimethyl-
sulphine bromide. This was afterwards, in 1875, acknowledged by
Cahours to be the case (Compt. rend., 80, 1317; also Ann. Chim.
Phys. [5], 10, 13), and he fully confirmed Dehn’s statement as to the
formation of diethylene disulphide and triethyl(or methy])sulphine
bromide as the main products of the reaction. It is necessary to
mention these facts, as descriptions of the purely imaginary salt,
(CH;),(C,H,)S.Br., occur in some of our dictionaries of chemistry.

According to Dehn, the two equations given above do not represent
the whole reaction; for in addition to the triethylsulphine bromide
there are formed the bromides of two other sulphine bases, but in
much smaller quantity than the former. His complete scheme for
the reaction includes the equations—

III. C.H,Br, + (C.H;).S as (C.H,)(C.H;).SBr.,
IV. C.H,Br, + C.H,S — (C.H,).SBrz.

That is to say, there are formed side by side three distinct sulphine
salts, one of which is of the ordinary kind containing tetrad sulphur,
whilst the others are of a different character and contain hexad
sulphur. In his second paper (Ber., 2, 479), Dehn states that he
regards the formation of the salt, (C,.H,)(C.2H;).SBr., as proved by his
experiments, and that of (C,H,),SBr, as not absolutely proved, but
shown to be very probable. Cahours, in his later papers, corrobo-
rated Dehn as to the formation in small quantity of at least one other
sulphine compound besides triethylsulphine bromide. He did not
himself investigate this product, but he expressed a doubt as to the

* Dehn wrote the formula O,H,S, but his description of the substance shows it
to have been the disulphide.

tee. 1 ee

nn een

CONTAINING THE ETHYLENE RADICAL. 251

accuracy of Dehn’s view of its constitution, a view, however, which
he most strangely misrepresented.

As regards equation 1V, one thing is at once obvious, that it
involves a neglect of the fact that it is diethylene disulphide, (C,H,),8,,
and not a substance of the formula C,H,S, that, according to Dehn’s
own showing, results from the primary reaction. Other considerations,
however, combine to make it difficult to accept Dehn’s theory con-
cerning these secondary products. It is based upon the idea that a
class of sulphinic salts can exist, resembling the sulphinows salts in
their general characters, and standing to them in a relation similar to
that of sulphuric anhydride to sulphurous anhydride. This view,
though plausible enough in 1865, is considerably weakened by the
fact that no additional evidence of its truth has been published since
that date. Indeed, if his theory be really a correct one, Dehn’s salts
stand alone, and occupy a position of very considerable importance to
the student of sulphur compounds. This was recognised by Divers,
who, in a recent paper bearing on the atomicity of sulphur (Trans.,
1885, 213), referred to Dehn’s salts, and to the halogen addition
products of trimethylsulpbine (described by Dobbin and myself) as
exceptions—and apparently the only ones—te his rule that sulphur is
never more than tetrad. The latter exception is, as yet, at any rate,
a somewhat doubtful one; and I believe that the sequel will show
that Dehn’s salts need form no exception at all.

Dehn was unable to isolate and examine the bromides themselves ;
and a perusal of his paper shows that his theory of their constitution
rests entirely on the properties and partial analyses of their platini-
chlorides. The question to be answered therefore resolves itself into
this :—Is there sufficient reason for believing that these salts were
respectively (C,H,)(C.H;).S"Ch,PtCl, and (C,H,),S"Cl,,PtCh, as he
represented them, or is it possible to substitute for these formule
others more in keeping with the experimental evidence and with
general probability ?

His method of procedure was as follows :—The cold dilute aqueous
solution of the mixed sulphine bromides, freed from diethylene
disulphide and ethyl bromide, was shaken with recently prepared
silver oxide, and the resulting mixture of hydroxides was treated
with excess of hydrochloric acid, warmed, and fractionally precipi-
tated with solution of platinic chloride. The triethylsulphine salt,
being soluble in hot water, was the last to form. The first salts
precipitated were yellow powders, apparently crystalline, practically
insoluble in cold or hot water and in dilute acids, but somewhat
soluble in strong aqueous ammonia. They were not acted on by
hydrogen sulphide in the cold, but were decomposed by a stream of
the gas when suspended in hot water.

252 MASSON ON SULPHINE SALTS

The first two fractions of this insoluble precipitate contained 39°13
and 38°39 per cent. of platinum. Dehn’s formula, (C,H,),SCl,,PtCl,,
requires 39°31 per cent. (taking Ptas 195). There was not sufficient
material for further analysis. The subsequent fractions had a
constant composition, which, as tested by determination of the
carbon, hydrogen,and platinum, agreed fairly well with that calculated
from the formula (C,H,)(C,H;),SCk,PtCh.

As this is all the actual evidence brought forward by Dehn in proof
of his theory of the existence of the sulphinic salts, that theory must
fall if it can be shown that there is a priori ground for believing that
these platinum salts had altogether different formule, and also that
Dehn’s own analyses agree as well, or better, with these formule than
they do with his own.

As regards the first point. The properties of Dehn’s platinum salts
are strikingly unlike those of most platinichlorides, but are identical
with those of the salt (C,H,).S.(CH;)Cl,PtCl,, described in the last
paper (p. 245) except in so far as “apparently crystalline” differs
from “amorphous.” This salt, like “Dehn’s, is an insoluble yellow
powder thrown down when a hot solution containing the parent
sulphine chloride is precipitated with platinic chloride. The iodide
corresponding to it is formed by the union of diethylene disulphide
(C,H,).S, with methyl iodide. It is not unnatural to suppose that the
same sulphide will unite, though less energetically and completely,
with ethyl bromide and even with ethylene dibromide. If these
analogous reactions can occur, they will give rise to sulphine salts of

the formule—

C,H, C.H;s C.H, C,H, C.Hy
S<on.>S<pr 94 S<¢.4,>5<prpr>S<¢,4,7*:
Diethylenesulphide-ethyl- Diethylenesulphide-ethylene-
sulphine bromide. disulphine bromide.

In Dehn’s reaction, as we have seen, diethylene disulphide is formed
as a product of the first change: it is formed together with ethyl
bromide and in the presence of ethylene bromide, and amidst
conditions of temperature, &c., specially devised to favour such
combinations. If, therefore, the salts whose formule are given above
are capable of being produced, they would certainly be formed in
Dehn’s experiment. The platinum salts would then have the
formule (C,H,),S,(C,H;)Cl,PtCl, and [(C,H,)2S, ]2(C.2H,)Clh,2PtCk,.
The following table gives a comparison of the percentage composition
required by the first of these with that required by Dehn’s theory and
with the numbers deduced from his analysis :—

CONTAINING THE ETHYLENE RADICAL. 253

Calculated for Found by Calculated for
(C.H4)28,(C.H;) Cl, PtCl,. Dehn. (C,H) (C,H) 28Cl,,PtCl,.
14°17 13°69
2°66 2°66
— 6°08
37°29 and 37°24 37:07
—_ 40°50

100°00

Though the differences are comparatively small, they are significant ;
and they do not favour Dehn’s formula. Had he estimated either
sulphur or chlorine, it is obvious that he would have come to a
different conclusion.

The evidence as to the constitution of the other sulphine salt, or
indeed of its existence, is so slight that it seems almost superfluous
to suggest that it was most probably {(C.H,).S,}.(C2H,)Cl.,2PtCl,.
This requires 38°50 per cent. of platinum: Dehn found in his first
fractions 39°13 and 38°39 per cent. But there is reason to believe,
from the following experiment, that the corresponding bromide, as
well as the salt (C,H,)S,(C,H;)Br, must have been produced in
Dehn’s reaction, though in small quantity.

Two sealed tubes were prepared, the one containing diethylene
disulphide with excess of ethyl bromide and the other diethylene
disulphide with excess of ethylene bromide. After being heated for
several hours at a temperature of 125—150°, they were cooled and
opened. In each case, free hydrobromic acid and disagreeable volatile
sulphur compounds were found to have been formed, but the bulk of
the disulphide was unaffected. By extraction with warm water,
however, small quantities of a sulphine bromide were obtained
from each tube; and these possessed the following characteristic
properties, a comparison of which with those of the salts of
diethylenesulphide-methyl-sulphine removes any possible doubt as to
their true nature.

(l.) They both undergo partial decomposition when their solutions
are evaporated to dryness, a white insoluble substance being
produced. The decomposition occurs to a greater extent than in the
case of the corresponding methy] salts.

(2.) They are both decomposed by heating with alkalis: a sudden
turbidity occurs, due to the separation of a white flocculent solid, and
a volatile sulphur compound is evolved having a strong and peculiar
odour. This odour is different in the two cases; for, whilst that
derived from the ethylsulphine exactly resembles the smell of the oil
CHS; (p. 247), that from the ethylene compound is of a more
pleasant character,

254 SNAPE: CERTAIN AROMATIC CYANATES

(3.) With recently precipitated silver oxide, they give alkaline
solutions, which, on the addition of hydrochloric acid, and, after
warming, of platinic chloride, yield insoluble yellow amorphous
precipitates.

These facts and the considerations already adduced seem to justify
the following equations, III and IV, being written in lieu of Dehn’s,
The whole scheme for the reaction becomes—

I. 2C,H,Br, + 2(C.sH),S = (C:H,).8, + 4C,H,Br.
II. C,H,Be + (C,H;).8 = (C,H;),SBr.
ILf. C,HsBr + (C.H,).S, = (C2H,):S.(C.H,)Br.
IV. C,H,Br, + 2(C,H,).S: = [(C2H,)2S2].(C2H,) Br.

It is probable that the iodides of ethyl and ethylene would combine
with diethylene disulphide much more readily and completely than
the bromides. If this beso, the products may repay an investigation,
more particularly the ethylene one. I propose therefore to continue
my experiments.

XXVIII.—Certain Aromatic Cyanates and Carbamates.
By H. Luioyp Snape, B.Sc.

CoNTINUING some experiments on phenyl cyanate already reported
(Chem. Soc. Trans., 1885, 770), I examined its action on pyridine.
No reaction takes place in the cold, when these liquids are mixed in
about molecular proportions, but on boiling, and then allowing the
mixture to remain in a vessel closed by a drying tube containing
calcium chloride, beautiful crystals are formed. These are only
sparingly soluble in ether, but somewhat more soluble in alcohol; on
recrystallisation from the latter, small quadratic plates are obtained
melting at 175°. This melting point, the form of the crystals, the
relation to solvents, and the fact that a large proportion of the pyri-
dine remains unchanged, suggested the probability that the phenyl
cyanate had condensed in the presence of this base to dipheny] di-
cyanate. On analysis this proved to be the case.
Calculated for
C,.H,yN20 Caleulated for
(i.e., CSHH;N + C,H;NCO). C\ 4H yo N20x. Found.
11°76 12°27

Many years ago Hofmann showed (Ann., Suppl. 1, 57; Ber., 3,
765, and 4, 246) that a small quantity of triethylphosphine brought

AND CARBAMATES. 255

about this change; but pyridine is more suitable as being much
cheaper and more readily obtained.

The account of the preparation of a phenylene dicyanate by Gatter-
mann and Wrampelmeyer (Ber., 18, 2604), by a method analogous to
that by which Hentschel obtained phenyl cyanate (Ber., 17, 1284),
suggested to me that other dicyanates might be prepared in a similar
manner, and I could then examine their action upon alcohols, as I
had previously done with phenyl cyanate. The action of phosgene on
benzidine was first investigated.

Diphenylene Diisocyanate, CO : N-C.H,yC.H,N : CO.

A current of phosgene was passed over well-dried benzidine hydro-
chloride contained in a fractional distillation flask, heated in a paraffin-
bath at 230° to 250°. A small quantity of a colourless oil passed
over, and solidified in the delivery tube leading from the flask. In
later experiments it was found preferable, as giving a better yield, to
couduct phosgene for about two hours over the salt heated in benzoic
acid vapour (b. p. 250°), finally distilling over the cyanate, still in
an atmosphere of phosgene, by means of a bath of diphenylamine
vapour (b. p. 310°). Probably a substituted carbamide is first formed,
as with aniline; and a higher temperature is required to completely
decompose this and to drive off the cyanate. The two phases of
reaction would, according to this hypothesis, be represented by the
following equations :—

(1.) NH,C,H,;C,HyNH, + COC], = 2HCl + 00< san

2.) COCNE CHS + COCL = CO:N-C.H:N:CO + 2HCL.
6 4

The distillate was almost pure and solidified in splendid long
needles. These were quite insoluble in cold water and in hydrochloric
acid, and therefore did not contain any of the original salt or free
benzidine. On the other hand, traces of hydrochloric acid adhered
somewhat persistently to the cyanate. On heating it, the greater part
melted and then sublimed, but a portion was always decomposed.
The vapour had the peculiar odour of phenyl cyanate. It softened a
little under 100°, and the purest looking specimen I obtained melted
at 122°. Another portion after sublimation, and in external appear-
ance almost as pure, melted at 116° to 120°. The difficulty of
obtaining it absolutely free from hydrochloric acid made it difficult to
fix within a few degrees the melting point of the pure substance. It
was soluble in ether.. Analysis :—

SNAPE: CERTAIN AROMATIC CYANATES

Calculated for
I

11:37

As was to be expected, this cyanate, like phenyl cyanate and
phenylene dicyanate, yielded carbamates with alcohol and phenol.

Diphenylene-diurethane or Diethyl Diphenylenedicarbamate,
COOC,H;NH-C,.H,-C,H,yNH-COOC,H;.

A portion of the above cyanate was dissolved in just as much
boiling alcohol as was necessary, the boiling continued for a few
minutes, and then the solution was allowed to remain undisturbed to
cool. The crystalline deposit was almost pure, and, when recrystallised
from alcohol after previous boiling with animal charcoal, separated in
beautiful feathers. These crystals melted about 226°, but it is some-
what difficult to determine exactly the melting point, as a decom-
position sets in at about the same temperature.

Calculated for
Found.
8°66

The constitution of this compound was further confirmed by pre-
paring it from ethyl chlorocarbonate and benzidine. Benzidine was
treated with the ethyl chlorocarbonate—an excess of the latter being
present—in a vessel supplied with a reflux condenser, and heated for
about 20 minutes in an oil-bath at 130°. On merely mixing, a
considerable amount of heat was developed, and the reaction then
mainly took place. After distilling off the excess of ethyl chloro-
carbonate, the remaining brownish-white solid was washed well with
water, and recrystallised from alcohol. The same peculiar feathery
crystals were obtained. These melted at about 230°. The slight
difference in the melting point is explained by the already-mentioned
difficulty of determination. The reaction in this case is as follows :—

2(NH,C,H,-C.H,-N H,) + 2(Cl-CO-OC,H;)=(NH,°C,HyC,Hy NH2,2HC!)
+ ( COOC,.H,;"N H ‘C;HyC.Hy N H ‘C OOC,H;) e

The benzidine hydrochloride which is simultaneously formed is sepa-
rated by its solubility in water.

Diphenyl Diphenylenedicarbamate,
COOC,H;'NH:C,H a C,HyNH-COOC,H;.

Diphenyl dicyanate was heated with a slight excess of phenol in an
oil-bath at 140°. These substances at first melted, and afterwards

AND CARBAMATES. 257

solidified again, showing the formation of a compound of higher
melting point than either. This was found to be but sparingly
soluble in most ordinary solvents, such as alcohol, ether, benzene, &c.
After previously washing well with ether to remove phenol, it was
recrystallised from glacial acetic acid, as this solvent seemed to be the
best. A large quantity of acetic acid was required. After a second
recrystallisation, the carbamate was obtained in beautiful crystalline
tables. These were washed with ether many times to remove acetic
acid, dried, and finally submitted to analysis.

Calculated for

Cog HopN 04. Found.

6°84

The melting point was about 240°, but was difficult to determine with
precision, as the substance decomposes with evolution of gas at about
the same temperature.

Metatoluylenediamine, CH,;C,H;(NH:)., was next treated with
phosgene: metatoluylene diisocyanate, C(H;(CH;( NCO), = [1:2:4],
was thus obtained. Toluylenediamine hydrochloride was treated with
phosgene in an exactly similar manner to that described for obtaining
diphenyl dicyanate, viz., successively in benzoic acid and diphenyl-
amine vapour-baths. The toluylene dicyanate passed over as a pale-
yellow liquid, which solidified to groups of needles almost perfectly
white. The needles were smaller than those of diphenylene dicyanate.
The compound as thus collected was almost pure, although con-
taminated to a slight extent with hydrochloric acid, which could not
be readily removed. The melting point was 94°. On distilling the
substance, it lost hydrochloric acid, and though it passed over mainly
unchanged, was in part decomposed, the melting point being con-
siderably lowered.

Calculated for
Found.

15°56

It was soluble in ether, and possessed the strong and unpleasant
odour of phenyl cyanate, especially noticeable on gently heating.

Thus, in all its properties this compound corresponds to, and is
evidently identical with, that described by Lussi (Ber., 7, 1263, and
8,291). He obtained in the first place the corresponding urethane,
by treating toluylenediamine with ethyl chlorocarbonate, and then
distilled this urethane with phosphoric anhydride. In order to con-
firm the identity, a portion of the cyanate, as prepared by me, was
boiled with alcohol, and on allowing it to stand, white silky needles
Separated. These were soluble in ether, and melted at 135°. These

VOL. XLIX. T

258 SNAPE: CERTAIN AROMATIC. CYANATES

properties agree with the description of the urethane given by Lassi.
For complete control, a small quantity was prepared as described by
him from ethyl chlorocarbonate and metatoluylenediamine, recrys-
tallising the product from alcohol. I thus obtained exactly similar
lustrous noedles, melting at 135°; they gave the following results
on analysis :—

Calculated for

Cy3H N20. Found.

Bi wcccccece coccscee 10°53 10°64

Lussi states the melting point to be 137°. I converted another por-
tion of the cyanate into the phenol carbamate.

Diphenyl-toluylene Dicarbamate, C,H;(CH;)(NH*COOC,H;).=
[1:2:4].

Metatoluylene isocyanate, as prepared above, was heated with an
excess of phenol in an oil-bath at from 130° to 150° for half an hour.
A gelatinous mass was obtained, which when boiled with ether to
remove phenol, became changed into a gray powder. This was re-
crystallised from hot glacial acetic acid in which it readily dissolved.
Clusters of white needles separated on cooling. These were collected,
well washed with ether to remove excess of acetic acid, and dried at
105°. The compound was slightly soluble in alcohol and ether and
melted at 147°5°.

Calculated for
Ca, HygN204. Found.

7°87

The next attempt which was made was to treat orthotoluylene-
diamine with phosgene. I did not succeed, however, in obtaining any
cyanate in this case. It was recently noticed by Gattermann and
Wrampelmeyer, that whilst para- and meta-phenylenediamine are
each converted respectively by phosgene into the corresponding
cyanates, orthophenylenediamine is not. Their experience and my
own may probably be explained by the suggestion that the ortho-
diamine is first converted into an ortho-substituted carbamide, and
that then the NH-groups are held so firmly by the CO-group as not
to permit of further decomposition by phosgene. Thus, in the above
case, the reaction is probably—

C.HACH)<Nie + COCh = 2HCI + CH,(CH)<N};>C0.

But the compound refuses to be further converted into the corre-
sponding dicyanate, as is the case when the NH-groups are attached
tv carbon-atoms further apart in the benzene-ring.

AND CARBAMATES. 259

The corresponding urethanes which would result from the acticn
of alcohol on these cyanates, can, however, be prepared by the action
of ethyl chlorocarbonate on the amines.

Orthophenylene-diwrethane, CsH,(NH:-COOC.H;), = [1:2], was
prepared in this way. Orthophenylenediamine was heated with an
excess of ethyl chlorocarbonate at 130°, the unaltered ethyl chloro-
carbonate distilled off, and the residue washed with water and hydro-
chloric acid. By recrystallisation from alcohol, after previously
boiling the solution with charcoal, lustrous needles melting at 88°
were obtained. The dried compound was submitted to analysis.

Calculated for ;
CoH igN 204. Found.
11-11 11°31

Orthotoluylenediamine also reacted with ethyl chlorocarbonate, and
crystals separated from the alcoholic extract of the product. I have
not as yet purified and analysed this compound.

Lastly phenylhydrazine was treated with phosgene. The latter was
conducted over phenylhydrazine hydrochloride, heated at 220° to
230° in a paraffin-bath. Towards the close of the operation, the tem-
perature was allowed to rise to 250°. A pale-brown oil distilled over,
and solidified on cooling to a pasty mass. The product of the reaction
possessed the powerful odour of phenyl cyanate, and the vapour
strongly attacked the eyes. On heating it in a fractional distillation
flask, hydrogen chloride was evolved, and a portion sublimed in
splendid long needles. Pressed out on porcelain, it gradually but
entirely sank into the plate. Instead of the reaction taking place as
I hoped, according to the equation—

C,H,NH-NH, + COC], = C,H;NH-NCO + 2HCl,

a solid combination of phenyl cyanate and hydrochloric acid had been
formed. That phenyl cyanate has the property of dissolving large
quantities of hydrogen chloride to form a solid mass is mentioned by
Hentschel (Ber., 17, 1285, and 18, 1178). That this curious com-
bination had taken place in this case was shown by the evolution of
hydrogen chloride on heating the product ; by the continued character-
istic and unpleasant odour of phenyl cyanate, which was so unbearable
that it could not be worked with for longer than a few minutes con-
secutively, and lastly by its conversion into carbanilide. This was
effected by simply boiling with water in a flask provided with a
reflux condenser until the odour of phenyl cyanate had disappeared.
The crystalline needles thus formed, after recrystallisation from
alcohol, were shown to have the same melting point and percentage of
nitrogen as carbanilide. Phenyl cyanate, as is well known, is readily
T 2

260 PICKERING: THE INFLUENCE OF TEMPERATURE

converted into carbanilide by the action of water. The experiment
did not allow me to determine whether phenylhydrazine was at first
decomposed, yielding aniline, which, together with phosgene gave
phenyl cyanate, or whether the cyanate, C,H, NH-NCO, I had ex-
pected to obtain was first formed, but then decomposed, yielding
phenyl cyanate.

I have much pleasure in taking this opportunity to express my best
thanks to Dr. Gattermann, for valuable counsel afforded by him to
me during the course of the work above described.

Chemical Laboratory,
University of Gottingen.

XXIX.—The Influence of Temperature on the Heat of Chemical
Combination.

By Spencer U. Picxertne, M.A., Professor of Chemistry at Bedford

College.

Our knowledge of the influence which temperature exerts on the
heat of chemical combination is at present of the most rudimentary
character. It has been established as one of the fundamental principles
of thermochemistry that the total heat evolved in the combination of
two substances at different temperatures, together with that necessary
to reduce the systems in the two cases to the same initial and final
temperatures, is a constant quantity—that where Q and Q’ repre-
sent the actual heat evolved during the combination of A with B at
T and T’, we shall have—
Q — Cy,(T’ — T) = Q — (Cy + Cp)(T’ — T),

C,, Cg, and Cy, being the specific heats of A, B, and the compound
which they form, respectively. Where the specific heat of the com-
pound is equal to the sum of those of its components, the heat which
these develop in combining will be a constant quantity independent
of temperature. The labours of Kopp have shown that such a con-
dition exists to a certain extent in the case of most solids, but we are
still in ignorance as to whether it is absolutely or only approximately
true, whether those variations which have in many cases been observed
in the specific heats of solids are regular in their nature, whether, in
fine, the compound AB at T is in every respect the same substance as
AB at T’.

ON THE HEAT OF CHEMICAL COMBINATION. 261

The present investigation shows that this is not the case. It will
appear that the relation between the constituent portions of the
molecule of a complex solid undergoes a series of modifications as the
temperature changes, the result of which is that the variations ex-
hibited in the heat of combination, and consequently in the specific
heat also, being conditioned by a different order of circumstances at
different temperatures, exhibit irregularities of a very marked cha-
racter.

As no methods have yet been devised for obtaining very accurate
measurements of specific heats through small, or indeed through
considerable ranges of temperature, the heat of combination appeared
to be the only available means of investigating the subject; in order,
however, to invest the determinations of it with a degree of ac-
curacy which would promise any chance of success, it was necessary
to confine them within the limits of ordinary atmospheric tempera-
tures, a small range of 25° C. only being thus available. Substances
which would be most likely to exhibit any variation in the heat of
their formation within these narrow limits would be those in which
the energy of combination was small. Hydrated salts afford many
instances of such a nature, and, consequently, they were chosen as
the subjects of this investigation; the heat of hydration presents,
moreover, one other material recommendation for the present purpose,
in that the number of operations necessary for a determination of it
is very limited, and the nature of these is very simple.

Whatever may be the complexity of the reactions occurring when
a salt is dissolved in water, it is generally accepted that the difference
between its heat of dissolution in the anhydrous and hydrated state
gives the heat of combination of the salt with its water of crystallisa-
tion,* provided we first subtract from the second of the above-named
quantities the heat absorbed in the conversion of the solid water
present into liquid water, for every fact which bears on the question
supports the view that in a solid hydrated salt the water is present as
a solid, and in the solution of such a salt as a liquid.

The heat of fusion of ice at temperatures other than 0° never
having been determined experimentally, it must be calculated by
means of the specific heats of water and ice; the latter of these has
also never been determined for temperatures above 0°, and, conse-
quently, some doubt must be felt as to the correctness of the values
calculated for the heat of fusion. The present experiments them-
selves, however, will be found to afford considerable evidence as to
the accuracy of these calculated values, and, moreover, any error in
them will not affect the main conclusions drawn from this work,

* This can only be so if the heat of combination of the various atoms composing
the salt molecule is unaffected by the combination of that molecule with the water.

262 PICKERING: THE INFLUENCE OF TEMPERATURE

inasmuch as such errors would affect the inclination of the curves
only, and not their nature. The latent heat of fusion of 1. mol.,
or 17°96 grams, of water, |», has been calculated from the equation—

bn = (1423°33 + (¢ — c’)T) cal.,
in which the specific heat of water,
e = 17°96 + 0°00072T + 0°0000162T”,

and the specific heat of ice, c’ = 9052 + 00185T. The various
values of /,, which is of course a negative quantity, are represented
in Plate VII, and it will be seen that, for the purposes of the present
communication, this is practically a straight line.

The heat of combination of a salt with its water of crystallisation,
both being in the solid state, will be given by the equation—

Q=M, - (My + tl»),

in which M, and M, represent the observed heat of dissolution of
anhydrous and hydrated salt, and x the number of molecules of water
which the latter contains. If, as has hitherto been assumed, the
specific heat of a hydrated salt is equal to the sum of that of the
anhydrous salt and the water (reckoned as solid water), the value of
Q will be a constant quantity, independent of temperature.

The more important conclusions arrived at in this paper are based
on differences so small as to be nearly on a par with those attributable
to experimental error; it will, therefore, be necessary to enter con-
siderably into detail to show that they can by no means be attributed
to this source.

Apparatus and Method Employed.

The apparatus employed was Berthelot’s open calorimeter, and full
details as to the method of operation adopted in this research will be
found in Phil. Mag., 1886, 21, 324. It will be sufficient to mention
here that, except in some of the earlier experiments, and in some of
those at very low temperatures, the temperature of the air was kept
constant within about 0°2° of the initial temperature of the calori-
metric water. This precaution, which naturally rendered the ex-
periments very tedious, was necessary, in order to reduce the errors
inherent in the determination of the rate of cooling to a minimum,
and it also obviated any important correction having to be applied on
account of the temperature of the salt before dissolution not being
identical with that of the water in which it was dissolved.

The rate of cooling (or heating) was determined in every single
experiment, and the necessary correction applied, for it was ascertained

ON THE HEAT OF CHEMICAL COMBINATION. 263

that not in one case out of a hundred did the salt dissolve in a space
of time sufficiently short to render such a correction superfluous.

The general method of procedure was modified in the case of
hydrated sodium sulphate only. The adhesive nature of this salt
rendered it impossible to shake it out of the weighing-tube into the
calorimeter without causing an inordinate amount of splashing; the
following device was, consequently, adopted in one series of experi-
ments (those with sample No. 3095). The salt was weighed out into
a very short wide test-tube, of which the water equivalent was
known, and the tube with its contents plunged bodily into the water ;
a platinum wire fused into the bottom of the tube served as a means
of holding the tube inverted till its contents had fallen out. It was
found, however, that the presence of these tubes in the calorimeter
interfered with the agitation of the liquid to such an extent that, at
low temperatures, a very long time elapsed before all the salt dis-
solved. In the experiments with other samples, therefore, the salt
was weighed out into the ordinary test-tubes on feet, but instead of
being shaken out into the water, it was shaken out on to the flat
disc which formed the end of the agitator, and which for that
purpose was momentarily raised till flush with the level of the water.
All splashing was thus avoided. The raising of the pounder was
repeated in each interval, during which the rate of cooling was
measured, so that any error due to evaporation from its surface
during the process might be allowed for; no additional error appears
to have been caused by this operation, as the average variation of
duplicate experiments in this series is not greater than that in other
cases.

Thermometers Employed.

Two calorimetric thermometers were at first employed, No. 55,083,
with a range of 1—16°C., and No. 55,081, with a range of 10—25°;
subsequently another instrument, ranging from 5—21°, was used
(No. 56,916), and in a few experiments No. 55,080, a fellow instru-
ment to No. 55,081, as well as No. 55,084, with a range of 0—26°.
They were divided into about 38 arbitrary degrees, each of which
were subdivided into tenths, these subdivisions being 1°2 mm. apart,
and each equivalent to about 0°038°C. In reading them, a pocket
lens was employed, and the fractional portions estimated to the
twentieth of the marked divisions, one estimation unit, therefore,
representing 0°002°. With No. 55,084, these quantities must be
doubled, as the range in its case is twice as great as with the others.

These calorimetric thermometers were originally standardised by
comparison with a natural standard, but as the work progressed’
it became apparent that this method was insufficient to afford

264 PICKERING: THE INFLUENCE OF TEMPERATURE

the amount of accuracy required, since a small error in the
reading of the standard would cause a very appreciable want of
concordance in the various instruments. The following expe-
dient was therefore adopted :—All the thermometers were compared
with No. 55,084, the range of which embraced those of all the others,
and which was thus converted into an ‘“ intermediate standard ;” as
this instrument is about three and a half times as delicate as the
natural standard, a corresponding increase in the concordance of the
calorimetric instruments was thus attained. It is true that this
“intermediate standard’”’ itself had to be standardised by comparison
with the natural standard, but any error made in this operation would
affect all the calorimetric instruments to the same extent, and would
not impair their concordance inter se, which was here the consideration
of most importance. The results thus obtained were further modified
by comparing the calorimetric instruments with each other, this being
effected by comparison with No. 56,216, the scale of which overlapped
that of all the others to a considerable extent, twelve or more com-
parisons at different points being made with each instrument, a fall or
rise of about 170 mm. (2°3°) being taken in each observation.

At the same time that this re-standardising was effected, all the
instruments were calibrated afresh, and all the several hundred experi-
ments which had been completed were recalculated. Some details
as to these calibrations and standardisations will be found in the Phil.
Mag., 1886, 21, 180.

In spite of the very great pains which were taken to produce an
absolute concordanee between the various thermometers, the results
were not altogether satisfactory ; indeed a certain regularity which
appeared in the differences of the results obtained with different
instruments, seemed to indicate the existence of some unknown imper-
fection inherent in the nature of mercurial thermometers, which pre-
vented the attainment of such concordance throughout the scales.

A device was consequently adopted, by means of which not only
the same thermometer, but also the same part of the thermometer, was
used in every experiment, whatever the temperature might be. This
was effected by removing the requisite amount of mercury into the
upper chamber of the instrument by the application of heat, thus
adjusting its reading to any particular temperature, and rendering
two experiments at different temperatures as comparable with each
other as if they had been performed at the same temperature : in fact,
for the purposes of the present investigation, the errors of calibration
and standardisation were thus obviated entirely.* At the same time,
the delicacy of the instruments was increased six- or seven-fold ; in two
of them (Nos. 62,839 and 63,616) the total range of about 600 mm.

* A similar device has, I believe, been used by Crafts.

ON THE HEAT UF CHEMICAL COMBINATION. 265

represented 3°6° and 386° respectively, each estimation unit (0°05 mm.

in this case) being equivalent to 0:0003° only. A full description of

these thermometers, together with some peculiarities which they

exhibit, and the manner in which these must be guarded against, has

been given in the Phil. Mag., 1886, 21, 330.
Quantities Used. :

The relative proportions of salt and water used in all these experi-
ments was 1 mol. of a single salt to 400—420 mols. of water, the
actual quantities being ,4,ths of a gram-molecule to 600 c.c.

Chemists have generally preferred to take proportions of water
not exceeding 200H,O, but though such a proportion gives twice the
temperature disturbance, the larger quantity, 400H,O, would appear
to offer greater facilities for very accurate measurements in most
eases. The estimation of the rate of cooling is certainly that part of
a determination which admits of the largest error, and the employ-
ment of 200H,0 in these experiments would have increased the rate
of cooling beyond the limits within which that rate is practically
constant in the case of three of the salts, while in all the cases in which
anhydrous salts were dealt with, the larger proportion of salt would
have required twice the time pounding and stirring to have effected
its dissolution, and would thus have doubled the correction necessary
for cooling. Moreover, with the instruments employed, a rise or fall
of as much as 5°, which would have occurred in some cases had
200H,0 been used, would have limited very materially the range of
temperature available for the series of experiments, and would have
rendered it impossible to use such delicate thermometers as the last
two described.

Another inducement for taking the larger proportion of water will
be mentioned when the calculation of results is discussed.

With double salts, the proportion of salt was only one-half of that
used in the other cases.

Experimental Error.

As the question of experimental error in these experiments has
been fully discussed in the Phil. Mag., 1886, 21, 324, it will be
necessary to mention here the general results only.

One estimation unit (0°06 mm.), in the case of the less delicate
thermometers, represents 16 cal. in the molecular heat of dissolution
of a single salt, while with the two more delicate ones it (0°05 mm.)
represents 2°5 cal.

The experimental error where the former were used was found to
be + 18:86 cal. for every single experiment; of this + 5°56 cal. is

266 PICKERING: THE INFLUENCE OF TEMPERATURE

due to errors in the calibration and standardisation of the instruments,
leaving + 13°3 cal. as the error of each experiment, when duplicate
determinations at the same temperature, and performed with the same
instruments, are compared. Of this error, + 5°71 cal. have been
traced to error in reading the instruments, and + 7°59 cal. to error in
manipulation. In cases where two thermometers were used simul.
taneously in tite calorimeter, the manipulation error is increased to
+ 9°53.

With the two very delicate instruments, the total probable error of
each experiment was only + 7°42 cal., this being the error when com-
paring experiments performed, not only at the same temperatures,
but at different temperatures also.

In the case of the double salts, all the above numbers become

doubled.

Calculation of Results.

The salt before its introduction into the calorimeter being at very
nearly the same temperature as the water, the correction due to the
“salt temperature” was generally very small, rarely amounting to
one estimation figure of the thermometer. The specific heats from

which any such correction was calculated were taken as follows :—-

Na,SO,,10H,0, 0°320 ; Na,SO,, 0-229; K,S0,, 0-190 ; Li,SO,, 0-200 ;
CuS0,,5H,0, 0-300; CuSQ,, 0°135 ; MgSO,,7H,O, 0°380;
MgS0O,, 0:222; CuSO,,K,S0,,6H,0, 0°260; CuS0,,K.SO,, 0°168 ;
MgS0,,K,S0,,6H,O, 0264; MgSO,,K,S0,, 0°170.

Having applied this correction, as well as that due to cooling (or
heating) during the dissolution, the heat evolved was calculated from
the initial and final temperatures thus obtained, according to the
simplified method adopted by Berthelot (Méc. Chim.,1, 190), which
consists in reckoning the number of cubic centimetres of water ori-
ginally taken as so many grams with a specific heat of 1, and neglect-
ing the salt present. It is true that the specific heat of solutions
containing 1 mol. of the salt to 200H,O has been determined for some
of the substances here dealt with (Marignac, Arch. des Sc. de la Bibl.
de Genéve; 1876), and that the employment of solutions of that
strength would have allowed of the adoption of a more accurate
method of calculation in such cases; but even in the few instances
where it would have been possible, its application would have been
limited to the experiments at one temperature only, since the data for
calculating the specific heats at any temperature between 0° and 20°
are entirely wanting. '

Berthelot shows that in many cases his method introduces but very
trifling errors, and gives far more accurate results than are obtained

CN THE HEAT OF CHEMICAL COMBINATION, 267

by assuming the specific heat of dilute solutions to be unity. With
solutions as dilute even as those containing 200H,0, it is liable, how-
ever, in some cases to lead to serious errors, taking for instance copper
sulphate. 23°8 grams of the anhydrous salt would be dissolved in
602:05 c.c.* (weighing 600°5 grams) at 24°; the specific heat of such
a solution at this temperature is 0°95, and hence its water equivalent
will be (600°5 + 23°8)0°95, or 593°1 grams, whereas according to
Berthelot’s method, it would be taken as 602°05 grams, thus causing
an error of 7th. This error, however, is greatly reduced by using
more dilute solutions. From Marignac’s work on the specific heat
of saline solutions (loc. sup. cit.), we may conclude that the specific
heat of solutions of various sulphates differ from each other to a
smaller extent the more dilute they are; that with 400H,O their
specific heat will probably be very near 0°98°, and that a rise of tem-
perature of 1° will cause an increase of approximately 0°0001 in this
specific heat. Taking now two experiments with copper sulphate with
400H,O0 at 24° and 4° respectively. At 24°, the water taken would
have measured 602:05 ¢.c., and weighed 600°50 grams. The weight
of salt taken would be 11°9 grams. The water equivalent of the
solution would be (600°5 + 11°9)0°98 = 600°05 grams, whereas it
would have been taken as 602°05 grams, the error, therefore, being
shoth. At 4° the water taken would be 601°74 c.c., weighing
601°74 grams, and the true water equivalent (601°74 + 11°9)0°'978 =
600°14, that taken 601°74, the error being 25%,th. Such errors (and I
have taken an instance where it would appear to be larger than in
most cases) would often be within the limits of the experimental
errors, and, as will be seen, the conclusions drawn in this communi-
cation do not depend on the absolute accuracy of the numbers given,
but on the relative accuracy of experiments with the same salt at
different temperatures (the error in comparing together the above
experiments would be only (545 — #5 = )re'soth), and on a com-
parison of the heat of dissolution of the anhydrous and hydrated
salts at the same temperature. All the experimental data, however,
will be given in full, so that in the event of future work completing
our knowledge of the specific heats of saline solutions, these results
may be recalculated in a more correct manner,

Preparation of the Salts.

The method generally adopted in preparing the hydrated salts
consisted in recrystallising them very carefully two or three times,

* I take the volumes which would actually have been used in the present experi-
ments; these are the volumes delivered by the same glass vessel at the various
temperatures. No difference would be made in the present calculations were the
same volumes taken at every one of the temperatures.

268 PICKERING : THE INFLUENCE OF TEMPERATURE

powdering the crystals finely, and washing them with water. This
powdering and washing is very necessary in the case of all sulphates
containing copper, or any allied metal, since, owing to the dissocia-
tion of the salt in solution, any mother-liquor adhering to the crystals
will be acid, and this will cause serious errors to arise in the colori-
metric determinations. It is also preferable in such cases not to boil
the solutions of the salt, since a high temperature greatly increases
the amount of this dissociation.

The hydrated salts were invariably sifted, sometimes through a silk
sieve.

Considerable difficulty was at first experienced in prepaing these
salts with exactly the theoretical amount of water, owing to efflores-
cence. The most satisfactory results were obtained by determining
the percentage of water present in the crystallised salt which
had been imperfectly dried, and then allowing it to lose whatever
excess of water it was found to contain, by exposing weighed quanti-
ties of it to the air with constant stirring. It was ascertained that
even a large quantity of a salt which retained excess of moisture
could, by sifting and mixing, be obtained sufficiently homogeneous
throughout to render such a method trustworthy.

To take one instance out of many, a sample of hydrated sodium
sulphate weighing 1 kilo. gave on analysis 56°612, 56°611, and 56°596
per cent. of water, numbers which are very closely concordant,
although the sample contained a considerable amount of adhering
moisture, the theoretical percentage for Na,SO,,10H,0 being only
55°88.

The most accurate and indeed the only satisfactory means of
analysing most of the sulphates here dealt with, was found to be a
determination of the percentage of water present in them, using very
large quantities for each analysis. This method has already been
discussed in a previous paper (Trans., 1885, 100), and the analytical
results will not be given in the present communication, since, unless
the water percentages were found to correspond to the theoretical
amount within experimental error (within about 0°03 per cent.), the
preparation was rejected.

In the case of a hydrated salt, such as sodium sulphate which parts
with its water of crystallisation very easily, it was found necessary to
prevent the salt from caking in the bottle in which it was preserved,
by constant agitation, for if it cakes some of the water from the
crystals in one part of the bottle distils into another part which
happens to be slightly cooler, so that the sample is no longer homo-
geneous. A whole series of experiments was spoilt by neglecting this
precaution.

It was found best to preserve the anhydrous salts in small bottles

ON THE HEAT OF CHEMICAL COMBINATION, 269

or sealed tubes, each containing sufficient for only two or three
determinations, so that exposure to the air might be avoided as much
as possible ; in order, however, to make sure that the contents of
each bottle or tube were precisely the same as regards their state of
purity, it was necessary to put the whole of the sample into a big bottle
immediately after its preparation, keep it there with continual shaking
for several days, and then after heating it in the bottle itself toa
temperature considerably above 100°, to transfer it to the smaller
vessels.

The experiments were so arranged that different determinations at
the same temperature were made after long intervals of time, so
that any alteration in the salt, whether of a molecular or of accidental
nature, might be detected. In no case was any such molecular altera-
tion observed, and only in one instance (with a sample of copper
sulphate) was any difference detected, and that was due to accident.

Specimens Used.

The following are the preparations which were used in the various
experiments. For the sake of reference, I have retained the numbers
by which they were distinguished when prepared.

Hydrated Magnesium Sulphate, MgSO,,7H,O0 = 245°48.—This
salt when pure neither effloresces nor deliquesces on exposure to
air; its preparation, therefore, is very easy. Three different speci-
mens, Nos. 2688, 2756, and 3417, were used in the majority of the
experiments, as well as three others in a few cases.

Anhydrous Magnesium Sulphate, MgSO, = 119°76.—Samples 2718
and 2754 were prepared by heating the salt in small quantities in a
platinum dish below a red heat. No. 3092 was dehydrated at a
temperature of 250° only.

Hydrated Potassium Magnesium Sulphate, MgK.(SO,),,6H,O = 401°4.
—Prepared by cooling a hot saturated solution of the component
sulphates. The details of its preparation in a state of purity are
given in this vol., p. 8. ,

Four different samples were used.

Anhydrous Potassium Magnesium Sulphate, a-MgK,(SO,), = 293°64.—
Five preparations of this salt were made. The difficulty of obtaining
it in the anhydrous condition without inducing a change which con-
verts it into a second modification necessitated the employment of
samples retaining small quantities of water (except with No. 3039)
for which a correction had to be applied. No. 3432 had been some-
what overheated, and gave numbers uniformly 1113 cal. lower than
the other samples. This amount was added to experiments with it
for convenience in deducing a curve. The peculiarities of this salt,

30 £37

STOATEA

ree O narm a alia gl

we www ts ©

270 PICKERING: THE INFLUENCE OF TEMPERATURE

and those of the corresponding copper compound, formed the subject
matter of a communication to the Society, and full details as to most
of the specimens here used will be found in this vol., p. 1 et seq.

Hydrated Copper Sulphate, CuSO,,5H,O = 248°64.—One specimen
only of this salt was used, No. 2736.

Anhydrous Copper Sulphate, CuSO, = 158°84.—Eight different pre-
parations were used. The first one, No. 2729, having been kept in
one bottle, was found to have accidentally absorbed a certain amount
of moisture, which rendered the numbers yielded by it in the later
determinations uniformly 160 cal. lower than those given in the earlier
ones. This amount has consequently been added to the later experi-
ments with it, as indicated in Table III. Nos. 3094 and 3401 were
preserved in small bottles, and the remaining specimens in sealed
tubes. All the specimens were prepared at a temperature between
200° and 250°.

Hydrated Potassium Copper Sulphate, CaK,(SO,)2,6H,O0 = 440°78.—
Two specimens of this salt, prepared in the same manner as the corre-
sponding magnesium compound, were used; one of them, however,
in two experiments only.

Anhydrous Potassium Copper Sulphate, a-CuK,(SO,), = 333°02.—
The four different specimens used were prepared by dehydrating the
crystallised ‘salt at temperatures between 115° and 130° (vide supra).

Hydrated Sodium Sulphate, Na,SO,,10H,O0 = 321°41.—Two diffe-
rent specimens, Nos. 3095 and 3418, were prepared from totally
different samples of the commercially pure salt.

Anhydrous Sodium Sulphate, Na,SO, = 141:°81.—Seven preparations
of this salt were examined. No. 3036, prepared by drying the
effloresced hydrated salt over sulphuric acid. Nos. 2775 and 3052,
samples dried at 100°. No. 3081, dried at 150°. These were all
obtained from the same lot of the hydrated salt.

Owing to the excessive degree in which this salt cakes, it was
impossible to make any accurate determinations with these samples
below 10°. It was found, however, that the anhydrous sulphate pre-
cipitated by heating a saturated solution of the salt above 30°
dissolved without difficulty, and evolved the same amount of heat-as
the non-crystalline specimens prepared by efflorescence (see Trans.,
1884, 689). Three samples (Nos. 3098, 3407, and 3604) were prepared
in this manner from different lots of the hydrated salt.

Potassium Sulphate, K,SO, = 173°88.—Four different preparations
were made, the finely-powdered salt being sifted through silk, and
then heated for many hours at 200—250°. These specimens did not
yield absolutely concordant numbers for their heat of dissolution, due
no doubt to the temperature at which they were prepared not being
exactly the same (see Trans., 1885, 98). The various results were

ON THE HEAT OF CHEMICAL COMBINATION: 271

reduced to uniformity with those of No. 3603, by adding 18 cal. to
those with 2725, — 38 cal. to those with 3428, and — 63 cal. to those
with 3613.

Anhydrous Lithium Sulphate, Li,SO, = 109°84.—The experiments
with this salt do not claim the same amount of accuracy as those with
the other salts, as the specimens used were not specially purified.
Four samples were prepared by drying the hydrated salt at 200° till
constant; No. 3093 appears to give results 81 cal. above those of the
other samples, and this amount has consequently been subtracted in
Table XII, which contains the results.

A few experiments were performed with the monhydrated sulphates
of magnesium, copper, and lithium. Details of the preparation of
the four different samples of the first of these will be found in the
Trans., 1885, 100. The copper salt, of which one specimen was used,
was prepared by heating the crystallised sulphate at 190°.

Two different specimens of the lithium sulphate were used in the
different experiments with that salt.

Calorimetric Results. Explanation of the Tables (p. 291, et seq.).

All the determinations of the heat of dissolution are collected in
the accompanying tables.

The thermometers employed are designated for brevity sake as 80,
81, 83, 84, 16, 39, and 616 respectively.

wis the weight in vacuo of the salt taken, W the water equivalent
of the calorimeter and its contents, the actual water being 601°687 +
0-016T° C. cc. (reckoned as so many grams with sp. heat = 1)
together with that (if any) which the salt contained.

T is the temperature of the salt before its introduction into the
calorimeter, expressed in degrees of the same thermometer as that
employed in this latter.

t and ¢’ are the corrected initial and final temperatures of the
calorimetric liquid given in the arbitrary degrees of the thermo-
meters employed, the corrections include* that due to the tempera-
ture of the salt introduced, the cooling (or heating) during its
dissolution, the exposure of the column of mercury to the atmospheric
temperature, and the calibration correction.

M represents the molecular heat of dissolution of the salt as given

by the equation—
a re
(jo Ory we
M= ’

* According to the equation Go ae we ¢ being the specific heat of the salt.

272 PICKERING: THE INFLUENCE OF TEMPERATURE

in which M is the molecular weight of the salt, and « the mean value
of the thermometric degrees in degrees centigrade, these being for—

80 = 0°39633
81 = 0°38663
83 = 0°38046
84 = 0°70659
16 = 0°38549
39 = 0°06025
616 = 0°064161

The corrections which are applied to M in certain cases will be under-
stood from what has already been mentioned.

In the case of a hydrated salt, the observed heat of dissolution is
represented by m, M being reserved for the “true” heat of dissolu-
tion = m — 2lm, l» being the heat of fusion of 1 gram-mol. of ice.

The last column gives the initial temperature of the experiment in
degrees centigrade, T. Where the mean results to which this
temperature applies are deduced from experiments in which the
various initial temperatures differed by more than about 0°08°, the
centigrade equivalent of each is given in the column under ¢.

In most cases, the means are arithmetical means, but in a few
instances some determinations have beey allowed a smaller weight
than others. With hydrated sodium sulphate, experiments with
sample 3418 have been allowed twice the value of the others.*

Explanation of the Plates.

The calorimetric results are given diagrammatically in the accom-
panying plates; each of these plates contains three curves, one repre-
senting the apparent heat of dissolution of the hydrated salt, the
second, to which the numbers on the right hand side refer, its “ true ”
heat of dissolution, and the third, to which the numbers on the left
refer, the heat of dissolution of the anhydrous salt. The continuous
thick lines represent the mean results given in the tables, the others
will be explained shortly. The higher the position of a point in any
of these curves, the greater is the heat development or the less the
heat absorption which it represents.

From the manner in which the curves are arranged, the vertical
distance between any points in those representing the heat of dissolu-
tion of the salt in its hydrated and anhydrous conditions is
proportional to the heat of combination of the salt with its water of

* Some of the numbers given in these tables will exhibit slight differences from
those quoted in previous papers. The present ones are the more accurate, as they
depend on a more correct calculation of the scales of the thermometers, and on 4
larger number of experimeuts.

ON THE HEAT OF CHEMICAL COMBINATION. 273

crystallisation in the solid state. The values deduced for this latter
are represented separately in the curves at the foot of each plate.

Critical Examination of the Results.

The most striking and most important feature of these curves is un-
doubtedly their irregularity. In all of them, there appear more or less
sudden breaks, or else they consist of two or more independent curves.

Before discussing them, and the conclusions which may be drawn
from them, it will be necessary to show the means which have been
taken to make vertain that these irregularities cannot be attributed
to experimental error, an explanation which will but naturally
suggest itself, when it is observed that an appropriate distribution of
such an error of less than 100 cal. would cause most of these
irregularities to disappear, and that even duplicate experiments at the
same temperature will sometimes differ by nearly this amount.

The two thermometers 81 and 83 were alone employed in the first
experiments, and the portion of the curves where the irregularities
most generally occur, and where they were first noticed, coincides
unfortunately with the point where one instrument was substituted
for the other (at about 13—15°). The conclusion drawn at first was,
of course, that the non-concordance of the instruments was the
explanation of these irregularities. Many considerations, however,
render such an explanation highly improbable.

(1.) The temperature at which the instruments were changed
being notified by a bar across the curves, it w.ll be seen that, although
a sudden dip near this point is generally due to the experiments with
the highest portion of 83, it is not always so; with MgSO, the dip is
due to experiments with 81, while in the case of Na,SO,, Na,SO,10H,0,
and CuK.,(SO,)2, there is no such dip noticeable, and with both
MeK.(SO,). and MgK,(SO,)2,6H,O the point in question was investi-
gated with both the instruments, and both of them gave identical
results.

(2.) Were the irregularities in the curve belonging to the hydrate:|
salt due toerrors in the scale of the thermometers, irregularities of an
exactly opposite nature would present themselves at nearly the samu
point in the curve of the anhydrous salt, but the reverse is the case,
The depressions or elevations in the one are generally more or less
reproduced in the other. If we were to apply corrections to the
thermometers such as would reduce any one curve to uniformity, it
would only bring into greater prominence the irregularities of the
other curves. Besides which, the errors in the calibration curves
which would have to be supposed to exist would be out of all
proportion to the known magnitude of the probable error from this
cause (see Phil. Mag., 1886, 21, 181).

VOL. XLIX, U

fa 23 Piers Sg ree He eyes ~~0g

274 PICKERING: THE INFLUENCE OF TEMPERATURE

(3.) The fact that these depressions do not always occur at the
same temperature in the different cases, and that others at totally
different temperatures exist, would alone be sufficient to disprove
their being attributable to such a source.

However, that no proof should be wanting, another thermometer,
No. 16, in which this debatable point occurred in the middle of the
scale, was introduced, and fresh experiments performed with it.
With potassium magnesium sulphate, both hydrated and anhydrous,
a whole series were performed in which it was used simultaneously
with one of the other instruments; the results with 16 only are
(Plate III) represented by a thinner line, and bring out the peculi-
arities of the curve even more strongly than those with the other
thermometers, which form the dotted curve, and in which the indi-
vidual experiments are marked by crosses. The determinations at 15°
were repeated thrice with 16, and the mean result here depends on
as many as 8 observed values. For the sake of greater security,
the experiments with the anhydrous and hydrated salt at the same
temperature were performed within half an hour of each other.

With hydrated sodium sulphate (Plate V), 16 was also used,
and the results with it are indicated as in the foregoing case. Except
at 12°3°, where a considerable divergence unfortunately occurs, these
results are concordant with those of the other instruments to the most
minute details.

Experiments with this thermometer were also performed in the case
of hydrated copper sulphate, hydrated magnesium sulphate, and
potassium sulphate, and agreed fully with those with the other
instruments, as will be seen from the tables, but their number was not
sufficient to make it worth while representing them separately in the
Plates.

In spite of these results, I did not yet feel satisfied, so remarkable
did the facts appear to me; and, consequently, started fresh series of
experiments with two instruments, 80 and 82, companion ones to 81
and 83, together with a third new one, 84, embracing all the
temperatures required ; what experiments were done with them were
confined chiefly to potassium sulphate, and will be found to be
thoroughly concordant with the others. But before the work with
these had progressed very far, the idea occurred of using the same
portion of the same instrument, and that a very much more delicate
one, for every experiment. All doubt as to the irregularities being
due to imperfections in the thermometers. were thus set at rest
effectually. This instrument, 39, which has been described above,
was applied in the case of hydrated magnesium sulphate, anhydrous
copper sulphate, potassium sulphate, and anhydrous sodium sulphate
(Plates I, IT, V, and V1) ; except with the latter salt, the results with

ON THE HEAT OF CHEMICAL COMBINATION. 275

it only are represented separately by thin lines; the dotted curves
representing those with the other instruments, and as will be seen
they bear out in every case the conclusions which had been arrived at
with the others, the results with 83 would appear to be generally
rather high in comparison with those of 39, but the main features of
the two are identical. The dotted curve given with copper sulphate
refers to three experiments with No. 616.

Such, then, are the means I have taken for assuring myself of the
correctness of these observations. But it is impossible by a mere
description in outline, or by the synoptical view presented by experi-
ments massed together in tables or diagrams, to convey any adequate
conception of the strength of this assurance: of how the few experi-
ments which first had exhibited any irregularity were repeated over
and over again, after long intervals of time, with different instruments
and with different samples; of how often the work, after being laid
aside as complete, would slowly assume the appearance of a phantom
conjured up by imperfections in the instruments, and of how often,
having been begun all over again, such an idea would be dispelled,
and conviction in the accuracy of the results enforced with double
power. And now, it is only after having fully investigated all the
imperfections of thermometers themselves, having obtained some more
delicate and more perfect instruments than any. yet produced, and
having devised a method of using them which obviates the greater
part of the experimental error, that I venture to bring the results
before this Society.

Hydrates in Solution.

A cursory inspection of these curves will show that many of them
must be still very imperfect. The changes in them are often confined
within such narrow limits that, but for the chance selection of some
particular temperature for an experiment, this change would never
have been detected. The number of observations which would be
necessary to establish satisfactorily the precise form of all the lines
would be enormous; and it must, therefore, be remembered that, with
the materials at present available, numerous as are the complications
exhibited, they may in reality be far more so.

Considering, in the first place, what conclusions may be drawn from
the examination of any one curve by itself, it will be seen that in no
case is the heat of dissolution, whether of a hydrated or anhydrous
salt, capable of being represented by a straight line, as is generally
assumed (see Berthelot, Ann. Chim. Phys., 29,310), nor even by a single
curve (except perhaps in the case of Na,SO, and CuK,(S0O,)2, though
probably not so in either of these cases). With some of the an-

{
i
a
4,
i"

a
‘Sim

a *

SRR

Ole SF

apes? Sas

276 PICKERING: THE INFLUENCE OF TEMPERATURE

hydrous salts taken, this discontinuity might be attributed to varia.
tion in their heat of combination with those molecules of water with
which they are associated in the solid form, but with hydrated salts no
such explanation is possible, and we -will take these, therefore, as pre-
senting the more simple problem.

At certain points, then, in these curves there is a more or less
sudden increase in the heat developed, and it seems impossible to
accoant for’such an increase in any way-except by the formation of
some new and higher hydrate.

What is generally termed the “hydrate” theory of solution may
probably be incapable of giving a full explanation of all the pheno.
mena of dissolution, but ‘t is, I think, proved by overwhelming
evidence that many salts in solution do exist in a state of combina-
tien with water, and it is highly probable that the hydrates which are
there present are often of a higher order than those known in an
isolated or solid form; that such is the case with the so-called
hydracids and alkalis, Berthelot’s researches on the heat developed
when they are diluted with varying amounts of water have shown,
and they have also shown that the same substance may form more
than one such hydrate under different conditions, whilst under inter-
mediate conditions these different hydrates coexist in a state of
partial dissociation. The same state of things would appear to hold
good in this case as regards salts, the agent which determines their
formation being heat instead of excess of water. From a study of any
of these curves, it would seem that as the temperature rises the amount
of the particular hydrate present reaches a maximum at a certain
point and then diminishes, thus producing a curve, when a further rise
induces a different state of equilibrium, and another hydrate begins to
form causing a fresh development of heat.*

This explanation, the only one which seems possible, involving the
idea that an increase of temperature favours a higher combination
between a salt and water, will at first sight appear to be opposed to
our general ideas on the subject ; but it must he remembered that we
ought to have no general ideas on this subject at all, for we have no
general knowledge on which to found them. We know, in the case
of sodium sulphate and a few other salts, that a certain temperature,

* Instead of the amount of the hydrate, or the number of water molecules con-
stituting it, being the variable quantity, it might be the energy with which the water
in any one hydrate is combined with the salt. But this I consider to be improbable,
for, as will be shown below, the temperatures at which an increase in the heat of dis-
solution of a salt occurs, are those at which a decrease in the energy of combination
of the salt with its water of crystallisation takes place; moreover, any increase in
this latter quantity is very small in comparison with the general increase in the heat
of dissolution of a salt as the temperature rises, and could not afford any explana-

tion of it.

eae

ON THE HEAT OF CHEMICAL COMBINATION. 277

generally a high one with reference to the salt in question, favours
the dissociation of the salt from the water with which it is combined
while in solution, or rather, that the balance of the many forces which
the temperature influences results in the dissociation of the hydrate :
but whether this be so in all cases, or whether it holds good also for the
temperatures far below the melting point of the salt or the boiling
point of solution, we know not, and.in the: face of this absence of
all evidence to the contrary, the present experiments must be accepted
as proof. A striking point in favour of such a view.is found in the
case of sodium sulphate itself, where, instead_of having any increased
development of heat at about 14° and 23°, as in other cases, none
occurs after 8°, that is, at a temperature 26° below that at which the
affinity of the salt for water is known to diminish ; the minor irregu-
larities which occur in this curve at higher temperatures being
probably due to experimental errors.

One case which was investigated with great care was that of
potassiam sulphate, owing to its being an anhydrous.salt in every
condition as far as is known, and one which, judging by the large
absorption of heat on its dissolution, possesses but small affinity for
water. As many as 58 experiments,. many of them with two thermo-
meters, were made; the delicate instrument 39 was used in 18
cases, and the results with it fully confirmed the previous ones, both
as to the existence of a dip at. 14°, where eight separate determina-
tions were made, and as to the curvature of theJine above and below
this point (Plate VI). Only at 19° are the experiments not alto-
gether concordant, and it may be that another dip occurs here, the
curvature between this point downwards to 14°.being more marked
so as to pass through the experiment at 16°9°. However this may be,
the existence of at least two curves and therefore two different
hydrated conditions seems certain. (The results with thermometer
39, below about 13°, give results uniformly lower than those with the
other thermometers, as has already been noticed.) Side by side with
the potassium sulphate curve in Plate VI is given that of lithium
sulphate, as well as a reproduction of that of anhydrous sodium
sulphate. These three curves will be seen to exhibit no stronger
resemblances to each other than they do.to those of less closely allied
salts. Lithium sulphate, like the potassium salt, shows a dip at#bout
14°, but in other respects the two cases are dissimilar, especially at
the lower temperatures, while sodium sulphate differs from both of.
them, the portion at 14°, and indeed that from 12—17°, being some-
what higher, rather than lower, than the rest of the curve. It will
have been seen that a considerable number of experiments were per-
formed in the case of this salt, Table 1X.

The experiment with lithium sulphate at 23° is not in accordance

278 PICKERING: THE INFLUENCE OF TEMPERATURE

with the rest of the curve ; for this, a sample of the salt was employed
which was not used in any other experiments, and it may have differed
from the other samples in purity; none of the experiments, indeed,
with this salt can be regarded as being of such an accurate nature
as those with the other salts (see p. 271).

No conclusions can be drawn as to the universality of a dip in the
curves at about 14°. It occurs in most of the present cases, it is
true, but it must be remembered that we are dealing with salts all
belonging to the same class, and that in the one which differs most
from its congeners—sodium sulphate—such a dip as we have just said
is absent; the precise point also at which it occurs is not identical in
every case, and the peculiarities which the various curves exhibit at
other temperatures are quite suflicient to differentiate them.

The salt which exhibits the greatest number of changes, and one in
which these changes have been well established, is hydrated potassium
magnesium sulphate (PlateIII). The heat of dissolution of this salt
between 2° and 24° constitutes no less than three complete curves,
together with the beginning of a fourth ene, the temperatures where
minima heat developments are observed being 2° (perhaps lower
in this case), 7°5°, 15°, and 243°. The increase which occurs after
the latter temperature caused the case of magnesium sulphate to be
investigated at these higher temperatures also, and an analogous rise
was discovered, beginning, however, at a lower point, about 22°. But
magnesium sulphate, unlike its double salt, exhibits no minimum point
at 7°5°, the heat developed from 14° down to 3° forming but one con-
tinuous curve. It should be remarked that the same confidence
cannot be placed in the experiments at very low temperatures as in
the other cases, the great difficulty of keeping the temperature of the
laboratory constant within.a.few degrees of zero, and the great effect
which the presence of the operator has on the temperature of the
calorimeter, rendering the operations less trustworthy.

The curves formed by the copper salts (Plates II and IV) differ
very considerably from those of the magnesium salts; the changes
which occur are much more abrupt in their nature; those of
hydrated copper sulphate, and of the hydrated double sulphate would
appear to consist of a series of straight lines rather than curves, and
the true heat of dissolution of hydrated copper sulphate is affected by
temperature to a considerably smaller extent than is magnesinm
sulphate, or indeed any of the other salts examined. It must be
ncticed, however, that, with the exception of anhydrous copper
sulphate, these copper salts were not investigated so fully as the
niagnesium ones.

It is somewhat remarkable that in no case does any marked change
appear at 4°, the temperature of the maximum density of water.

—" — —~ — _ a | rp" am

ON THE HEAT OF CHEMICAL COMBINATION.

It is curious that the extent to which the apparent heat of dissolu-
tion of all three of the hydrated single sulphates varies with the
temperature is very small, it being in some cases practically constant
for a considerable range of temperature. This, however, is not so
with the double sulphates; it would appear as if the potassium
sulphate which is present in them still retained its individuality to a
considerable extent, its heat of dissolution when combined in the
double salt being affected in the same way, though not to the same
extent, as when in the free state.

The true “ molecular” nature of these double salts is also brought
into prominence by the fact that in order to render the curves of their
heats of dissolution analogous to those of the single salts, we must, as
has been done in these tables, represent the heat evolved on half the
scale only. We would aimost appear to be dealing with two molecules
instead of one, and thus getting a double rate of variation as the
temperature falls or rises.*

It should also be remarked that what is here called the “true heat
of dissolution” of both hydrated copper sulphate and hydrated mag-
nesium sulphate is a positive quantity of no mean value, and that such
a quantity even does not represent what is in reality the “true” heat
of dissolution of the hydrated salt; for from it should be subtracted
the heat of fusion of the salt itself, at present an unknown quantity.
If the heat thus evolved be due to the combination of the already
hydrated salt with more water—and it is difficult to find any other
explanation of such an evolution—either this additional water must
be considerable in amount, or else the energy with which it is com-
bined must be great, and bear no small ratio to that with which the
water of crystallisation is combined.

The series of curves referring to any particular salt are of too rough
a character, and their precise meaning too imperfectly understood, to
render it profitable to draw further conclusions from them; but with
more complete data and more perfect knowledge, it would evidently
be possible by prolonging them to obtain numeric values of the heat
of formation of each fresh hydrate, and it would appear possible, if
not probable, that this heat would not bear any very inconsiderable
ratio to that of the formation of the known hydrates, it being thus
unnecessary to imagine that the new ones are of a much more com-
plicated character than those already known.

The occurrence of these successive changes in the constitution of a
salt in solution, involving successive developments of heat, will
involve also successive variations in the specific heat of the solution,

* This is in reality a consequence of the heat of formation of these salts being

affected to only a small extent by temperature, and hence also of their specific heats
being practically equal to the sum of those of the constituent sulphates (see p. 288).

280 PICKERING: THE INFLUENCE OF TEMPERATURE

too small no doubt to be rendered apparent by the imperfect means
which we possess at present for its direct estimation, but none the less
certain notwithstanding. We thus have a curious picture of the
temperature of a saline solution rising under the influence of a constant
source of heat, not with a regular motion, but in a series of pulsations,
recurring at irregular intervals, often considerably less than 10°

apart.
Water of Crystallisation.

We now pass to a consideration of those conclusions which may be
drawn from a comparison of the curves of a salt in the hydrated and
anhydrous condition. The difference between the quantities which they
represent give the heat of combination of the salt with its water of
crystallisation; this,it must be clearly understood, refers simply to
the known solid hydrate, and is entirely independent of the com-
binations existing in solution.

The results are given in the lower diagrams in the various plates.
An inspection of these will show that the heat of this combination is
not a constant quantity, it does not even vary in any regular manner,
but resembles in nature the heat of dissolution of a salt.

There appears in the first place to be a general tendency towards
an increase in the energy with which the salt and water are combined,
and such a tendency is in full accordance with the fact already
observed of higher hydrates being formed in solution as the tempe-
rature is increased. It may be objected that here at any rate
we know that a rise of temperature favours loss of water with a
solid hydrated salt; but this knowledge, if it can be called such, is
derived from experiments at considerably higher temperatures than
wny here dealt with, and, as far as I know, in none of these observa-
tions has the effect of temperature only, or the influence of
temperature on the one factor, with elimination of its effects on
other concomitant factors, been ascertained. When a salt is heated
in air under ordinary conditions, it may be the increased avidity for
moisture with which the air becomes endowed, and not any dis-
sociating influence of the temperature itself, which causes the salt to
part with its water. Or the loss may occur not because the energy
with which it is combined is diminished, but because other tendencies,
those towards different arrangements of the atoms in the molecule, for
instance, are increased. In the present experiments such possibilities,
however, are eliminated, since their results depend on the difference
between two quantities which in every respect but one are precisely
similar; in every case we have identical initial and final states; such
experiments, therefore, are the only ones which can be admitted as
evidence on the subject, and the evidence which they bear is not, I

ON THE HEAT OF CHEMICAL COMBINATION. 281

think, much opposed to our general and unauthorised notions on the
question, when we remember the lowness of the temperatures with
which we are dealing, and when we notice that in the case of
sodium sulphate, an increase in these temperatures, low as regards a
thermometer, but high as regards the salt, causes a rapid full instead
of rise in the energy of the combination after about 15°.

In addition to all this, it must be pointed out that we are but
observing in the case of these so-called molecular compounds a con-
dition which has been well established in the case of many atomic
compounds. A sufficiently high temperature will dissociate the con-
stituents of water; but, nevertheless, the heat of their combination
(the water being in the gaseous state) increases with the temperature,
or does so at any rate at all those temperatures where we have any
knowledge of the specific heats of the substances concerned. It is
quite possible, and, indeed, unless some great changes in specific heats
occur, must often be, that, even when a substance is actually partially
dissociated, the heat of combination of its constituents in those com-
pound molecules still existing, is greater than it is at lower tempe-
ratures when no dissociation exists; the heat and the energy of the
combination of the constituents being, as a matter of fact, but one of
the many factors which determine this dissociation.

It must be borne in mind that the variations in the heat of combi-
nation of the salt with its water cannot be looked upon as being so
accurately determined as those of the heat of dissolutiou of the
hydrated salt, for they depend on the values of two quantities instead
of one. It would be rash to state that the break at 14° in the case of
MgSO,,7H,0 for instance (Plate I), is not due to experimental error ;
it is but faintly marked in the mean curve, and by taking the different
results of the different thermometers its appearance is very much
altered; moreover, the number of determinations with the anhydrous
salt were not large.

With copper sulphate (Plate II), however, it is otherwise. The
heat of combination between 2° and 23° is represented by two complete
curves and a portion of a third; the curves are steep, and throughout
there is a tendency towards a rapid increase in the heat of combination.
The first curve, that extending from 2° to 9° (and to a certain extent
the second curve also), will be seen to be due chiefly to the peculiarities
exhibited in the heat of dissolution of the anhydrous salt, and, in spite
of the large number of determinations which were made between 4”
and 9° with this salt, I cannot feel altogether satisfied as to it. The
results with the delicate thermometer 39 (the thin line) reduce con-
siderably the curvature at 6° which was given by the other instruments
(the thick dotted line); but still it is well marked, and it should be
remarked that an extra precaution was generally taken here of using

282 PICKERING: THE INFLUENCE OF TEMPERATURE

the salt contained in any one of the sealed tubes in which it was kept
at two or three different temperatures, and performing the experi-
ments consecutively. Three experiments performed with the most
perfect thermometer (616, the thin dotted line) exhibited no curva-
ture at all, but they are inexplicable by the side of the others, and the
general inclination of the curve. Of the existence of a second curve
from 9—17°, and a portion of a third one from 17—23°, there can,
however, be no doubt. A marked similarity also appears between
the various curves ; allowing for experimental irregularities, the incli-
nation of the rising and falling portions of each is the same, while a
gradual increase in extent prevails throughout them.

Coming now to the double salts, it will be found that similar
irregularities exist in the case of potassium magnesium sulphate
(Plate III). The heat of combination is represented by one curve
from 4—11°, by another less abrupt one from 11° to about 19°, from
which point a third curve would seem to start; the results with
thermometer 16 are given in the thin line side by side with those
of the other instruments (the dotted line), and it will be seen that
while they exaggerate some of the features, they make others less
apparent, and show clearly the futility of insisting on every minute
detail in the curves, and the necessity for considering their general
bearings only.

With this salt, however, and also with sodium sulphate, more
weight may be attached to the results at present under discussion
than in the case of the other salts, for the curves of both the hydrated
and anhydrous salts were investigated with the same degree of care.
With the magnesium potassium sulphate the extra precaution was
adopted in one series of experiments of effecting the determinations
with the hydrated and anhydrous compound as nearly as possible at
the same time.

With potassium copper sulphate (Plate IV) we have a line of a
totally different nature to that of the corresponding magnesium salt ;
a short sharp curve exists between 13° and 15°, which would be looked
upon with suspicion, were it not that this temperature evidently forms
a starting point for a different order of conditions; the inclination of
the line being totally different after it from what it is before. The
straightness of the two portions of the line is remarkable, as also is
the abruptness of the change; this abruptness appears throughout to
be a characteristic of the copper salts in contradistinction to those of
magnesium.

It should not be forgotten that the scale on which these curves of
the double sulphates are represented is only half that used for the
single salt, their peculiarities are therefore diminished. They will be
found in Plate VII on the larger scale.

ON THE HEAT OF CHEMICAL COMBINATION.

The only other case to mention is that of sodium sulphate (Plate V),
and here we have a curve of a very different character from those of
the other salts. There is a sharp increase in the heat of combination
from 5° to 8° followed by as sharp a decline, another slight rise
occurs afterwards as far as 13°; after this, the heat remains constant
for the next two or three degrees, and then declines regularly to the
end, instead of rising as it does in most of the other cases. Such I
interpret the nature of this curve to be: if we take the results of
thermometer 16 at 12°, where there is unfortunately some error with
one of the instruments, and use it on all the details exhibited at
higher temperatures, we should get the excrescence at 8°, repeated in
a less marked manner at 12°, and again more feebly still at 16°5°,
with a shadow of a third repetition at 21°; butalthough the results of
the different instruments are remarkably concordant (except at 12°),
the variations causing the irregularities at the higher temperatures
are so slight that they may be attributed to experimental error,
though, it is true, the number of experiments at each point is con-
siderable ; at any rate it is safer not to insist on their actual existence.

The curve of the anhydrous salt was examined with special care in
the case of sodium sulphate in order to ascertain whether it exhibited
any peculiarities at those temperatures where the hydrated salt did so,
but no such resemblance even of the faintest description could be
traced anywhere. Indeed, the only indication which exists of the
anhydrous salt not giving a continuous curve throughout is that the
portion between 12° and 17° is decidedly higher than the rest of the line.

The experiments with the monohydrated sulphates quoted in
Table XIII are too few in number to give much weight to any
conclusions drawn from them, but, such as they are, they tend to
show that the same irregularities exist in the heat of formation of the
lower hydrates as in that of the higher ones. The copper salt, how-
ever, is the only one where the temperature was varied sufficiently
for such purposes.

The following are the values which may be deduced in the various
cases :—

(1.) For CuSO, + H,O (solid)—

At 13°00°, 5615 cal.; at 17°83°, 5743 cal.; and at 22°88°, 5914 cal.,

numbers showing a rapid increase in the heat of combination with
the temperature, as with the pentahydrated salt, though not at a
regular, but at an increasing rate, this rate being 26°5 cal. per
degree between the first and second temperature, and 33°8 between
the second and third.
2.) For MgSO, + H,0 (solid)—
7012 cal. at 22°3°.

284 PICKERING : THE INFLUENCE OF TEMPERATURE

(3.) For Li,SO, + H,0 (solid) —
At 13°8°, 1267 cal.; and at 12°97’, 1417 cal. Mean, 1342 cal. at
13°39° ;
to this latter, however, no very special value is to be attached (see
p. 271).

Here, then, we have a series of changes analogous to those taking
place in saline solutions, but more remarkable, inasmuch as the
substances which exhibit them undergo no change in centesimal
composition.

That changes of temperature bring about changes in the constiiu-
tion, not only of complex molecules such as those of hydrated salts,
but of the more simple bodies which are termed atomic compounds,
as well as of the elements themselves, is already well known in some
few cases; but the universality of these changes, and the frequency of
their occurrence, is a totally new feature.

Our ideas of the stability of a compound must be considerably
modified ; we must recognise that the substance which we are dealing
with at one temperature may be totally different in constitution from
the same substance at a temperature only 10° higher or lower. These
changes can only be attributed to alterations in the relations between
the various atoms or molecules in the hydrated salt. They cannot be
explained by changes in the grouping of the many molecules which
(probably) act as physical units, for the nature of the determina-
tions eliminates entirely the influence of any such changes. These
variations in the heat of combination of the components of the
molecule will, as in the case of solutions, necessitate slight corre-
sponding changes in the specific heat of the substance, and we shall
consequently have the temperature of the salt under a constant source
of heat, rising in an undulatory instead of a regular manner.

One other very important feature remains to be noticed. On
referring to the plates, it will be seen that the high portions in those
curves which represent the heat of combination of the salt with its
water of crystallisation, coincide with the low portions in those repre-
senting the heat developed in the dissolution of that salt, and vice
versa; the stronger the energy with which the few molecules of
water in the solid salt are retained, the smaller is the number of
molecules in the liquid hydrate. It would appear as if a rise of
temperature causes an increase in the energy of the combination of
the water with the salt up to certain point, but that a further
increase, instead of raising this energy still higher, takes a different
direction, and endows the salt with a capacity for taking up a larger
number of molecules; but inasmuch as the amount of water held is
now greater, each individual portion of it is held with diminished

285

ON THE HEAT OF CHEMICAL COMBINATION.

energy. Such would appear to be the case, but it is not possible to
affirm it with confidence, for we are not studying the salt in the
same state in both cases; in one case it is in the solid, and in the
other in the liquid condition; but another instance of a like nature,
not open to such objections, will be brought forward shortly, and the
strong presumptive evidence which we have here for supposing the
relations between certain portions of a molecule to be influenced by
a further addition to it, is one which is certainly in accordance with
our general ideas on “atomic” compounds, and one which I have
elsewhere (J. Chem. Soc., Abstr. of Proceedings, 1885-6, 112) argued
on other grounds to hold good with reference to water of hydration.

It will follow from the fact that the heat of combination of a salt
with its water is not a constant quantity, that the specific heat of the
hydrated salt is not exactly equal to the sum of those of the anhydrous
salt and water (solid) in it. This conclusion will hold good even if
we assume an error to have occurred in the number taken for the
specific heat of ice at these high temperatures (see p. 262), for, not to
mention the irregularities in the curves themselves, the application of
a correction in this latter,’such as would reduce the curve (i.e., that
of the heat of combination) to a horizontal line in one case, would
only increase its divergence from such a line in another (cf. the latter
portions of the copper sulphate and sodium sulphate curves).

The general tendency exhibited in the majority of cases being in
the direction of an increase of the heat of combination with a rise of
temperature, it follows (p. 260) that the specific heat of the hydrated
salt will be somewhat less than the sum of those of the solid water
and anhydrous salt. It is difficult, however, to speak with any
certainty of these “ general tendencies,” for the range of temperatures
in the present experiments is very small in comparison with that
through which the salt is capable of existing.

An inspection of the numbers given for the specific heats of salts,
and the great differences in those obtained by different observers, will
make it apparent why any direct determinations of these specific
heats should have failed in elucidating these facts.

It will be necessary to stop for a moment to enquire what effect a
more accurate method of calculation, and a more perfect knowledge
of the data on which those calculations were based, might have on
the general conclusions arrived at here.

To begin with, it is only by assuming sudden alterations in the
specific heat of an anhydrous salt, or of ice, that the alterations in the
curves here given could be eliminated. Such alterations would
involve similar conclusions as regards the constitution of the salt (or
of ice), as we have here drawn as to that of the hydrated salt, and
would thus be but a poor explanaticn. The fact that the curves of

286 PICKERING: THE INFLUENCE OF TEMPERATURE

the true heat of dissolution of the hydrated salt preserve a course so
nearly parallel to those of the anhydrous salts in most cases, and that
they converge to a considerable extent in the case of sodium sulphate
(the only case in which we have extraneous reasons for expecting
that they would converge, i.e., that a rise in the temperature would
diminish the energy with which the solid salt retains its water), is
strong evidence that the water in a crystallised salt is solid water, and
that in the dissolved salt is liquid water, and also that the numbers
which we have taken for the heat of fusion of this solid water are
very near the truth.

Secondly, what would be the general drift of the error caused by
taking the volume of the solution as being equivalent to so many
grams with a sp. heat of unity instead of basing the calculations
on the actual sp. heat of the solution. On p. 267, where copper
sulphate was taken as an instance, the error being greater in its case
than in any other, it was shown that, as far as we could judge from
the very imperfect data at our disposal, whatever the error in the
actual quantities might be, the relative error in the comparison of two
experiments at 4° and 24° respectively, would be but ,.,5th; the
sp. heat of the solution at the latter temperature by comparison
with the sp. heat at the former would be under-estimated by that
amount. Referring to Plate II, we see that 16700 cal. is the heat of
dissolution of the anhydrous salt as given at 24°, it therefore ought
to have been ;,55th greater, i.e., 16810 cal.; in the same way, the
apparent heat of dissolution of the hydrated salt should have been
— 2820 instead of — 2800, giving for its “ true” heat of dissolution 5340
instead of 5360; the correction of the calculations would thus give
(16810 — 5340 = ) 11470 cal., as the heat of combination of the solid
with its water, instead of (16700 — 5360 = ) 11340 cal. which was the
value taken. A more correct method of calculation would thus only
bring out into stronger prominence the increase in this latter quantity
as the temperature rises.

Lastly, these experiments themselves have shown the existence of a
more subtle source of error, and one which it is impossible to
allow for. The initial temperature of the water is that at which
the dissolution must be regarded as being effected, the final tem-
perature serving only as a means of measuring the heat developed;
but the proper use of this final temperature depends on our
knowledge of the sp. heat of the solution between it and the initial
temperature, and, owing to the variations here exhibited in this, it
will not be the same for the interval from ¢ to ¢’ when ¢’ is higher than
t, as it is for a like interval when ?’ is below #, i.e., when dissolving
the anhydrous and hydrated salts at the same initial temperatures we
shall generally be using liquids possessing different sp. heats in the

ON THE HEAT OF CHEMICAL COMBINATION, 287

two cases. Taking as an instance experiments at 16° with potassium
magnesium sulphate, and remembering that it is the curve of the
hydrated salt which gives a representation of the sp. heat of the
solution, this point 16° occurs where the curve is rising, 7.e., where an
increased development of heat and, therefore, a diminished sp.
heat, is brought about by an increase in temperature. In the experi-
ment here with the anhydrous salt, the rise measured from 16° to (say)
17° will have corresponded to a smaller development than the fail,
measured in the case of the hydrated salt, from 16° to 15°, the
former quantity will have been over-estimated, and the latter under-
estimated, the individual experiment should have been plotted at the
points marked A and B respectively ; these in each case will be lower
in the diagram, and, if the fall measured is of the same extent as the
rise measured, they will be lower to the same extent, thus making no
difference in the distance between them, 7.e., in the heat of the
combination of the salt with its water as deduced from them. In
other cases, where the rise and fall measured were not equal, an error
will have occurred from this cause which may have resulted in the
exaggeration or diminution of the peculiarities of this curve, but it
will probably have been of very small extent, and will have affected
the general nature of the curve but slightly.

Heat of Formation of Double Salis.

The experiments which have already been described afforded the
means for the determination of another quantity, namely, the
heat of combination of two sulphates to form the double sulphates,
MgK,(S0O,). and CuK,(SO,)2, in the solid state. Where z represents
this amount, and M the molecular heat of dissolution of the two single
and double sulphates—

@ = (My + My.) — Muex — N,

in which N represents the heat developed on mixing solutions of the
constituent sulphates. This latter quantity has been measured by
various physicists, and has been given by them as nil in the case of
both copper sulphate and magnesium sulphate. A few fresh deter-
minations of it are given in Table XIV. In order to render these
determinations more accurate, blank experiments were interspersed
with the others, and a small correction, which would appear to be
generally necessary, owing to loss or gain of heat while the solutions
were being mixed, was thus made.

The results bear out fully the conclusions of former experimenters as
to the value of N being negligible, and consequently the heat of com-
bination of the two sulphates in the double salt will be the difference
between the heat of dissolution of the anhydrous double salt, and the

288 PICKERING: THE INFLUENCE OF TEMPERATURE

sum of those of its constituents in the anhydrous state also. The
various values thus deduced for these quantities at different tempera-
tures are plotted out into curves in Plate VII.

The general features shown in the case of the magnesium salt is
that of two curves meeting at about 14° where the heat of com.
bination is at a minimum; the portion between 2° and 6° would
appear to form part of another curve; but in this case, more than in
previous ones, we must not insist too stringently on the details which
the lines exhibit, as each point in them is dependent on the correct
estimation of three separate quantities, each of which is of consider-
able magnitude.

With the copper salt, we find a minimum reached at about the
same temperature, 14°, the heat of combination decreasing gradually,
and perhaps regularly, up to this point, after which it again increases,
but at a very much greater rate.*

Here then, with the components of a double salt, we have the same
remarkable variations in the energy of combination as we have
between a salt and its water of crystallisation, and similar conclusions
may be drawn from them.

Side by side with these curves are given those which represent the
heat of combination of the same double salt with its water of hydra-
tion (drawn to the same scale as the former), and a comparison of the
two leads to important conclusions. It will be seen that, with no
exceptions other than what may be attributed to experimental error,
the temperatures at which the one quantity attains a maximum is
the temperature at which the other attains a minimum. This is very
strikingly displayed in the cases of both salts, and fully bears out the
conclusion which was drawn ubove, that the relationship between any
two parts of a complex molecule is influenced by any change in the
other parts of it. Here we appear to have a principle of compensa-
tion carried out between the various parts of the solid molecule, just
as in the former case an increase of affinity exhibited by a salt in
taking up a further supply of water, seemed to leave it but a smaller
amount to be expended on that already attached to it.

Although this principle of intramolecular compensation may be
more confidently affirmed from these experiments with the double
salts than it was with the hydrated salts, it can yet be regarded only
as strongly probable, and not absolutely proved : even here we are not
dealing with the same molecule in both cases, for the heat of hydration

* Owing to a clerical error in a previous communication (this vol., 7 and 15),
the heat of dissolution of copper sulphate was quoted as being 600 cal. less than it
should have been, and the heat of formation of the double salt was consequently
under-estimated to the same extent. The present numbers contain also some
further vorrections of an unimportant nature (see foot note, p. 272).

ON THE HEAT OF CHEMICAL COMBINATION. 289

is measured in the hydrated salt, whereas the heat of combination of
the other constituents is measured in the anhydrous salt. It is diffi-
cult to conceive, however, any means whereby a nearer approach to
positive evidence could be obtained.

It will be noticed that with both pairs of curves in Plate VII, the
one individual curve is determined from two quantities, and the other
from three quantities, and that of these quantities one only (the heat
of dissolution of the anhydrous salt) is common to the two; thus the
data on which these two curves are based are to a considerable extent
independent of each other.

Summary and Conclusion.

The main points of the present communication may be summarised
as follows :—

1. The heat of dissolution of a salt at different temperatures is not
represented by a straight line, but by a curve, or generally by a series
of curves.

2. Each of these curves probably indicates the formation and
existence of some fresh hydrate in the solution.

3. The formation of higher hydrates as the temperature rises would
not appear to hold good when we approach a temperature which is
near the melting point of the solid hydrated salt, as in the case of
sodium sulphate.

4. From the constant variations in the thermal phenomena attend-
ing dissolution at different temperatures, it follows that the specific
heats of saline solutions exhibit similar variations, remaining constant,
or even varying at a uniform rate, throughout very limited ranges
only.

5. The heat of combination of a salt with its water of crystallisation
in the solid state undergoes similar series of changes.

6. From this it follows that the specific heat of such a salt is not
a constant quantity, and that it is not exactly equal to the sum of
that of the anhydrous salt and the solid water present, but is generally
somewhat less.

7. The general tendency between 2° and 23° is towards an increase
inthe energy with which the water is combined.

8. Any temperature at which a higher hydrate is formed by a salt
in solution, is marked by a decrease in the heat of combination of the
salt with its water of crystallisation in the solid state.

9. The heat of combination of two sulphates to form a double salt
is subject to variations of the same nature as that of the combination
of a salt with water.

10. The variations in the heat of combination of the two salts is

VOL. XLIX. x

290 PICKERING: THE INFLUENCE OF TEMPERATURE

attended by variations of an opposite sign in the heat of combination
of the double salt and the water combined with it.

11. Most of the salts here investigated show some points of simi-
larity in the curves which they yield, but each possesses specific pecu-
liarities.

12. The peculiarities exhibited by a single salt appear to be repro-
duced to a certain extent in the double salts which they form, illus-
trated by the rise in the curves of both the magnesium salts at about
22°, and the character of abruptness displayed in the case of both the
copper salts.

In drawing these conclusions care must be taken that they be not
understood in too general a sense. The temperature of 14° appears,
for instance, to correspond so often to a depression in the curves, that
we might be led to attribute this depression to changes undergone by
the water molecules themselves at this temperature; but it must be
remembered that the salts which have been investigated belong not
only to the same type, the sulphates, but chiefly also to the same class
of that type, and that the one which differs most from the others in its
nature (sodium sulphate) exhibits a totally different behaviour at this
temperature; and even those salts which are most similar in their
nature exhibit very marked differences at many temperatures in the
heat of their dissolution.

It would indeed have been more satisfactory to have examined a
greater variety of salts, and this would have been done had not the
work been begun with a different object in view, and had it not pro-
gressed a considerable way towards its conclusion before its more
important bearings became apparent. It is my hope, however, to
extend the work at some future time, though the laborious and trying
nature of it may prevent my doing so for some time; no form of appa-
ratus has at present been devised whereby such experiments as these
would not suffer in accuracy by being performed at temperatures other
than that of the surrounding air, and the absolute dependence of one’s
work on the atmospheric conditions of a changeable climate, the
great waste of time which its changes entail, together with the hard-
ship of working for days together at temperatures but a few degrees
above zero, and on other occasions in an atmosphere superheated some
20° or 25° by burning gas, and the constant attention requisite in
maintaining these temperatures for a long period of time, render
the work of a most laborious and unenviable description.

Z
2)
Lomi
=
<
Zz
onl
a2)
a
e)
=)
4
a
>)
lo)
A
eS)
ise)
=)
Fa
°o
-
-
<3)
i=}
ea)
jen
Z
°

eLL- FT
6S- LT

“‘OZE- ZZ
°

OTPET
6ZP6T

TS00%
8866T
82006 89002

€1006
Soest { O8A#6T

i
68102

(48902)
Z6L0Z
EhL0Z

6220. 9
828-9
cie-9

cOZI- 9
61F-9

C69F- Z
ehs- 9
PPE-9
TIt-9

cc6s-9

998-9
TE9- 4
ch0-9
T1¢-9
$8¢-9
ce9-.9

10-8
89-2
(00. 9) £8-#T
(96-&) OL- ST
6L- SZ
(48- ST) 88-8
18-Z1
(W-ST) OL- ZT
¥Z- Sl

(PF. ST) 80. 1
(¢8. FI) &- 21
(S89- #1) Z6- 21
IZ: 61

08: ¢¢

OF- SE

SI- $8

€8
€8
€8
€8
€8
18
18
18
18
18
18
18
18
T8
18
18

DARA DDODARAROANDG

FSLZ

8TLzZ *

81Lz

PoLS

PSl%

BIZ ©

POLS
FSLS
8122
81LzZ
** 2608

"8 OTsz
seeeees GoIZ

POLS

pus SIlZzZ

eeeeee $OLZ

teeeeees OTyg

Haste Sr aa

4-4

“ULIOT TT,

‘oidmeg

1 Ov {92.611 ="OS2W ‘a7vydjing wniseuboyy snosphyuy fo uoynjossigy fo yoayT—'] a14aV J,

‘swp1b 726.8 = *70U
€

. oe ee ae tte

Haste orn aa

|
co
P
&
<
=
a
ral
i)
te
—)
~
Z
3
z
Lam
i
=
e

e-0l 4

PICKERING

‘swosb TTF.8T =
OF ;
€

°j0UuL

8F-26S = O'HL*OSSI ‘a7vyding wnysoubnyy poyosphynjideyy fo uoynjossig fo yoayT—J] aiavy,

292

7 OLIF—| 988.
vstor S&zb—-| S6T-
Zreol— GLO8-
GZIg.
88Z-
SZrot— 160.
062.
GLSE.-
€cZ.
Z81- TT €9T-
G0z.
CP6L-
GOPZ.
C6LI-
‘O68T-ST ZLL0I— SEZ -

. 8gZ.
ZSI-
G02.

CSL. & (e6-2) 611-¢ ) LZ-L19 | 69%-8T
coll -¥ (F0.€) ¢608-¢ . Il- 919 | 41%-9T
C68L- IT 160.1 . 92-2419 | LES-8T
C62P. 9T ZrL- LT . 29-419 | 968- ST

1F9- ST . 626-91 | Z8- 8e- L419 | 464-81

LL8-ST : R96- 9T LT | GG. L19 | 0ZS-ST

GZS. ST S19. 9T . ZI- L419 | $62-8T
CZ8P- ST OFS. OT , $e. $89 | 924-61
G63Z- FZ , GZIS-. sz | T8- og. 219 | 18Z2-8T

SBF: BE | (go. 17) _989-OF | 00- TH | 9-119 | FOF-8T
CE8I-2e . ess. OF | 09- OF 3g 69. 419 | PIS-8T
CLL. ZE | (SI8- ZT) ZL8-0F | 99. ZF | 09- £19 | 9TF-8T
668-28 | (p17 97) 9F9-0F | SF-0F | 68 69. 419 | SES-8T
C998: ZE . 9FS-OF | Lb. IF ¥9- L19 | FZF-8T

Zbz- LT = sI| — a

eee. 08 | (20°80) Jeo, | s\| 9 } 26.819 ZL0- ST

99-81 | , 188. 61
ggg. its (2g. $1)

See HF DODDDOH RAR A eee

IFO. £8 | 26. ze } 99. 819 | €0g- ZT

%,
ro)
Lol
=
<
7
Lond
=)
|
°
=)
>)
—
=
<2)
jen]
S)
Fes
.)
oo
=
&
jen
_
Zz
)

‘4m “ ‘ | ‘4 pooyT) "Mh

(‘panurnu0oo)—*JJ @IAV I,

Cesgct { eee OG-1& OST. EP | LL-OU) Ol6- LZ-&L | | *8-bLO rrs- Ll 6098
EZ9Sl 689-16 | 681-89 ™ OCP - LP-SL re-FlY 600-2 609€
PPOST CRSP. OF SB | 20-92 CO-F19 | 9FG-IL | PIPE
SCOST S1€S-08 | (9) oT) | OL-92 c0.F19 L06-1L | PIPE
9LOST G98S. 08 GZL¥-SZ | 14-92 c0-F19 886-Il | FIFE
18981 16¥- 08 GIPP-SZ | Zg-SzZ SO. P19 Ch6-IT | FIFE
ILLS1 889-18 GLES-93 | 92-92 80- $19 666-1 | Ste
66SST €89-18 | (g7 47) 199-98 | 99-92 | 80. P19 296-Il | elite
L69ST 9€8- TE | 682-92 | 8-2 80-719 | 186-IT | e1te
PZLST SOLF- IE | GL6E-9% | 61-82 | 80-b19 | 486-01 | eIre
GOSST = O9T + SFOST chog-Se | (89.21) 94z-08 | 96-82 | | OL-FI9 | 826-IL | 6222
OL8ST z9¢-c¢ | (09-21) SzZF-08 | 80-F8 | | 10-419 | 006-IL | 6222
{ TP6ST ccgz- FF | (€8-80 e¢8-IL | 46-8 |, 68: FLO 066-11 | 6098
6LS1 09-1 7) 619-6 _

| oeeat iF forse | 72-80) $10.86 | o0-88 j 86: S19 | 166-1T | TOV
£96C1 ZS1-88 (c6¢. 1) T10O-€& ZO- && OL-F19 Te6- IT TOFS

OL9ST

86981

J
ZS6S1
esesTt
ZIGST=O9T + ZSLST
E1091
S&1-ST| FE6ST mt na
68091
98-91} Z809T aneet
SII9T COLT:

<2)
2
5
e
<
pe
fl
=
s
B
fey
=)
=)
Oo
7
is
EB
Zz
Lomi
=

96- ZI) S2Z9T

.
1D 19 19 1 1 HH OF Ho MM FAO MOH MOM MW MWY

FAH OON wD

PICKERING

', = ° “CLOT T, ‘ejdureg

‘swv.b E1g.TT = ‘jou . [$8.9cL = Osu ‘aznydjng seddog snouphyuy fo uorynossig fo wayy—I] 214VJ,

294

SO8FT SsI8-8 99F-IT | (#¢-8) eZEe9.9 E.¢ 96-819 40-21 Site

20ST SOIb-& | SI00-21T | (16-8) 169-2 $L-8 €&-F19 006 -OT 10FE

PZ0ST | 988-08 | 646.2% | (0-6) EFI-zI | 20-0 IT-809 CF6- 11 119

6L0SI | 91-62 | ZIZ-eh| (SI-%) 920.51 | 6F- eT 1-809 *86-IL | Z19¢

9L0ST L02-% *8Z-FL | (8-6) 240-01 | 66-6 cg. F19 09¢-0L | 10PE

r reect { ShEST P16-% €89- LT (gg.¢) 694-81 | 18-1 00-#19 896-IL | F608

6ZEST=091T +69IST | OF8-F 999- LT 4 928-Z1 | ZP-ST 00-F19 O16-TL | 62Lz°

9Lzgt { Seset C86Z- 18 | PEL. eh (90.9) S0FS-8E | 14-21 8Z-F19 Lv0-Z1 | 6098

( 6LZS1T | 198-08 | 086. ZF 6II-ZL | &9-#T 8Z-F19 48-11 | 6098

8SZSI | 890-18 | 98¢-2F | (ZTO. 81G-IT | 82-9 8% -FI9 046-1 | T19¢

PLZS1 4 PSSST IpZSt | 68-1 | 9E4-eF | (9¢0. G68-Z1 | 94-2T #1-809 896-IT | I19¢

SSEST | 918-08 | Shz6.zb | (Lo. ¢809-21 | 68-TIT F1L-809 €6F-IL | T19¢e

6&ZST SII9-62 | ZIP- eh | (eZ. 006-81 | 82-8 OF - 809 Lv0-Z2E | Z19¢

Lerest { SceST 9996-0 | S&Z0- 61 LOL-®L | 22-81 00-419 | ¢zo.zt | zzre

O9ES1T=091 + 00ZSI | Sh98-% | SeEo. 6T ‘ OLI-FL | 92-#T 00-F19 G48-IL | 6zLz

Zz
i}
—
H
<
2
mm
[oat
=
o
oO
Hw
a
oO
=
=
&
td
oO
ics
=)
=
<i
&
se
&
A
°o

j
72 “a ‘4 | ULLOU ‘a “m | ‘atdurrg

(‘pan u1j}U09)—J II 414AVv

( eeger J syst

wien, . of . Se . tS
‘$9.e6% = 70S'*s*OS3IT -agpydyng wnisoubryy wnissvj0q snosphyup fo uornpossry fo wayT— A ATAVI,

FPS. L&8- SI-F19 | 498-41 | yeLz
ggg. eee. 3h. L8-¥19 | 461-81 | 9E2z

CP90. Zs0- < 19- #19 | FS8-8I | 9EZz
P26. GL88. . 1S- $19 | TL4-8T | 9E2z

cots. S66L- 08-19 | S9I-61 | 9E4z

C96 - GZZI- 9F- F19 | PET-8L | 98Lz
$90. ZL6- }

S6IZ. gzI. 08-919 | F9F-8T | 98Lz
162. Sh0Z } ; ; ,
GFLF. Ze |(SZL-8T) 9ZP-88 TS-9T9 | 664-81 | 9842
8ZPF- (cos. et) 738 #8 LS-¥19 | SG¢-8T | 98Lz °
99P - C968. 8S 09.19 | 489-81 | 9842
818. Z8L- &Z } ‘
C089. C865. ST ZS-919 | 9Z8-8L | 9822

, 608- 961- LZ }
ecge.o | e860. 86.91 1S-919 | 994-81 | 984z
SIZ8-0 | ¢Z4¢0.- 626- 62 \
298.0 | Tee. £89.61 ZL- 919 | PLS-8T | 9ELZ
ST98-0 OFS. IZ Z8-F19 | ZL-81T | 9ELz
118-0 ° TI9- 12 99-19 | 00S-8T | 9EZz
098-0 ° TSS. ZZ 6S-F19 | 9ZE-8T | 9ELz
TPS. 0 ° 6L0-1F }
928.0 128. 08 €9-919 | SOT-8T | 9842

98-0 . Z6L- && 69- P19 e082 9ELz

5 @

HN Oo WO

THE INFLUENCE OF TEMPERATURE

828-0 . S899. €€ €F- P19 | 80S-LZT | 98Lz

PICKERING

“19 ; =e 9 “UuMeq yy) «6M ‘oqdurag

‘swvib QF9.QT = “Jou "7 ‘$9.86 = O°HS*OS1D ‘agvydjng seddogp poezwuphynueg fo wonossyy fo yoaF]—' AT WAV,

296

z
o
onl
{sl
<
a
—
[ee]
|
)
5
=|
<
=)
—
:
=)
ics
oO
HH
<
&
8)
sal
7,
co)

‘082,92

€Z60T

£960T 8Z60T

PLEIL
TZPIT
Pe9TT
6ILIT
T661T
LZ0Z1

SsOZT

0Z0ZT

0296
8696
PPL
+696
8066
1866
LOFOT
66801

{oo
6I80T

io
I60TT
99601
0680T

ZOPTT
OSPIT

SOLTT
peLtl

{oat
POIZT
6S1Z1
68611
ShOZT
{oe
9L61L

2 +
Stitt
“48 +
is +
Stitt
eltitt+
Stitt
€tit+
elit +
€tIt+
Stttt
Slit +!
eltit+
eliIt+
€ltt+
etttt+
48 +
Slit +
€liIt+t
ettt+
48 +|
e€lttt+
€ltIt+
€liItt

86h8
6L8L
S&h6
TTS6
Tg98
T8¢s
c6L48
ST8s8
$626
9826
96°66
90L6
€rle
8466
€°86
LLL6
Z260T
T9LOL
6PS0T
L980T
LESTI
Z6S0T
TZ901
8Z80T

48
Ss

TLOZI
Se6II
OZ +|SZ6TT
LIb +| SP9LT
Lib +| 6SSTT
SII +| SLZPIT

+
+

ClLZ8- T
L8@-T
62eG-T
GLVS-T
99€- T
9L8-T
Szss-T
SPOP- T
O9P-T
8LP-T
L6G T
g9¢-T
Lé¢-T
$6¢-T
9s¢.T
S6E¢-T
soeZ-T
SZl9-T
999-T
ys9-T
ShZ8-T
199-T
919-T
TéZ-T
CL88- T
ST&6-T
P06. T
688. T
O16. T
Tss. T
SZEé8-T
GLP8. T

L1g-¢
228-6
¢862Z - YT
O€F- 9T
268-9
SOP: ST
c90P-6
PEP -ZS
964- ST
296. 8Z
C686 -6T
102-8
068 -%Z
€L1-88
P28 -¥ZS
$P09- PL
6806 -PT
STIT- 8ST
ZLS- IE
O€€- 1Z
C898- FZ
TO8- 9€
LSS. 92
06S -0€
OIL. FE
SOL. PE
PEL- PE
c1S6-€€

VES- BE
S80. FE
G66. EE

800 -6&

G68T-F
G&z-8
COLL- FT
GZ88- FL
920-P
Ze0- LT
20-8
C680- IZ
988: FT
PLP- LZ
CZ9P- ST
oP9- TE
SSE -Z
845-98

892 -8Z
(16-1) coo. et
(96- #1) SLT -ST

66P-9T

916- 6%

919-61

PPO. &Z

OFI- $8

TS88- ¥Z

698-82

G8ZZ- ZE

GI&Z-ZE
O&Z- ZE
CZ90- ZE

PZE- SE

GL8I- ZE

LIl- 28

GOUT. LE

(86. #1)

€9-2
T¥-8
8 -PL
0S- PI

88- 9T
GZ-06

98-22

8&-E
96-9

¥S-Z1
10: eI
G9. ST
L&- 02
EE: &Z
6L-¥Z
P8- 8Z
61- ZE
g¢- 18
| 4h. BE
| 00- $€
| 9§-9E
| 88-82
| 04-82
| $4.98

S6- $19
L6-€19
10. FL9
T0-P19

}¥2-s19
} a2-st9
} 18. s19
} es.sto
} 98.s19

} 88. e19

VI-PI9
81-19

| 26. st9
€Z- P19
} 26. st9

L6- S19
LZ- 19
LZ-¥19
LZ. 19
Lz-$19
LZ. F19
Lz. #19
LZ- 19
0€- P19

| 924-01 |oo0e Pur g6zz “FZ

91Z-TT
640-TI
2L1-II

6v0- TT
£66 -OT
196-0T
vOT-TT
600- TT

T20- TT

890- IT
LL0- Tt

89T- IT
680. IT
SP0- TI

68T. TT
L420. TT
OFL- IT
TLO- IT
670. TT
P8T- TT
¥60- IT
L90. TT
OvG- TT

teeeees Zoe
000€ PUB 96LS
000€ PUB 9626

reeeeee Zone
teeeeee Zone
*e °° SErE
"*** SSPE
29° zone

eevee ZEFE

000E PUB 96Lz
tteeees ZODe

*eeeeee ZEPrEs
000€ PUB 9622
seeeees core

eee eeee ZErs
teeeee* Bene
teeeees Bene
000 PUB 96LZ
sees es ONDE
veneer ’* O6L6
seeeees Bagge
*eeeeee rsos
‘ee eee ZErEe

"83
33
a id

“06

“6r

“L

*poyoor100

‘u0T}
-00110,)

“4

BUS Cling

“MM

‘ojdureg

‘$9.26 = "OS *M*OS 31

*smvib 2,

ToO-TT

€

qOUW

‘agpydjng wnisoubnyg wnissnj0g snosphyup fo uonunjpossyy fo 2vaFT— A

298 PICKERING: THE INFLUENCE OF TEMPERATURE

Taste VI.—Heat of Dissolution of Hexhydrated Potassium Magnesium

|
Sample. ‘ , Therm.

81 36°74 : 35°7375
81 31°08 8 30-618
81 31°66 “4 30 *7955
81 29°15 9°54: : 28-019
16 — ¢ 38 2725
81 28 84 *84é F 27-277
16 — . 37 °5285
81 24°76 ° 23 8425
16 — : 33-591
2789 5° S11: 81 23 49 3: 21-579
2789 ’ “( 81 20°63 “645 “8s 19 ‘O85

‘ 81 20-10 20° 18°712
| oe 16 — “53! 28 -942
81 15-93 “4 14.870
16 — 5 725 25 -090
81 14°97 “ht 13-514
81 14°87 133 13°526
81 13-21 ‘61: 11 -9785
16 — 3°81: 22-166
81 13°72 3°16 ; 11-511
81 13-16 “0: : 11 +4825
81 12°10 3°05 ; 11°443
16 — 3°24 ' 21-627
83 36°19 36: ; 34°980
16 — 3 3S 21°7685
83 32-08 31 °6: 29-917
16 — 16°7785
83 27°51 . 25°8175
16 — 12-7235
83 23°80 9175 22-182
83 20°06 "95: 138
16 — *8885 60935
83 16 *50 ‘0: 15° 262
- 5488 16 — 0 2°258
. 2789 ‘ . 83 13-25 3 °6465 11°826 |
. 2789 5° . 83 13°69 | 7 11-8815 |
. 3433 5° “¢ 83 12°88 606 (5°81) 10°7355 |
. 3433 5°23! “¢ 83 8-13 “014 6°144 |
. 2789 . 83 7°45 | 38°771 1°915 |

SoS PNP SF PF She

_
o

. 3433
. 2789

a =

. 2775
. 2789

. 2789
. 2789

. 3433

in
=
SE 3

—
te

I
~I

_
ie)
Pee ee ee eee ee eee ee ee ee ee el ee lm! lm! hClUmre.!hCUmre.!hLhCUmre

—_
~
. . ° “ ° a aa °
) @ or =
“OF — an YO a“ =a".

—
on

_
fr)
>
ou

. 3133

_
~I

~
@

. 3433
. 3433
. 2789
. 3433

ON THE HEAT OF CHEMICAL COMBINATION.

Sulphate.

MgSO,,K,S0,,6H,O = 401-40 ;

> 80

; mol. = 15:053 grams.

61m. M.

— 9788
—9769
—9849
— 9759
— 9900
— 9833
— 9875
— 9994

—9951
—9910
— 10026
— 10261
— 10325
— 10353
— 10274
— 10371
— 10417
— 10515
(—10454)
—10511
| —10554
—10411
— 10405
— 10601
— 10483
— 10741
— 10600

— 11059
— 11076
— 11291
— 11408
— 11363
— 11422
— 11447

1°
1°58
1°
1°5s
1°5
1°5
1°
1°5
1:
1°5
1°5
1‘!
1:
1°6
1°6
1°6
1°
1°6:
1°
1°6
1°5
1°6
1°63
1°6
1°6
1°7:
1°
1:
1°6
1°
1°
1°73
1:
17
1°
1°
1°
1°
1‘8

— 9804
— 9867

— 9968

}
-
f-
I
}-
I

}

~ 10394
} — 10490
} ~ 10533

} — 10408

——

—9227

— 9145
— 9076
— 9019

— 8941

— 8867

— 8854 |

— 8726 — 3149

6819 661-¢ c6II-F Ls-T c6- $19 ** OR ‘PELE
OIZL €28- 6 GLEI-8 Z0-6 L0-¥19 co°* Ogre
1891 ec9L 9¢8-FI (10.9) 60T-€T 1é-T 00- #19 ** "OR “PELE
1294 COCh- FL (G0.9) $LZ-8T GZ. ST 00-F19 "* 9ZbE

9PIg ion ° CEIZ- 0Z (IZ-8) S886. 8T 66-81 €0-F19 ** 9ZtS
Ze18 C601 -22 (¥6-8) 18-02 26-02 c0-F19 ‘OR ‘FELZ

0L06 819-88 1SZ- 20-28 Z1- 19 ‘ ** 9fFE

$606 4 9606 Z09- 88 GLZ1-ZE SI- 8 Z0- #19 ** QFLZ
9116 989-8 €91- 2 ZS: FES Z0-P19 ** HFLZ

PLE O16: &Z CELE. 2% G9 .1Z 1Z-19 “OR ‘PELE

L996 ZEG- 8S GOL. 12 ¥6- SZ 1Z- ¥19 awa

60L6 4 Z9L6 ee CRES- SZ S6LL- 12 £S. 92 12-19 “99° GHLZ
ShL6 SFG. ZZ L68- 12 GE. 0Z Z1-¥19 . "29° GELS

GO68- ZZ CPRE- 1Z ZS-81 Z1- 19 es ae

886- ZZ 9°F- 12 29-81 ZI-F19 sooo Ates
188- S8 L&l- 88 SI. 48 14Z- P19 "OP PPLE
LILZ pus

PIP. SE FI8- $8 09. 88 TLz-F19 SIS-ZL_ | ‘9bLz ‘PRLS

a Asdwornea

;
-
<j
fa
=
m4
fal
s
fa
3
a
oO
:
4
fe
Z
Lo
io
ee)

“4 + “m | ‘oidureg

‘swnib QQF.ZT = “Jom ™
‘zo.eee = "os*m*09uD ‘agvydjng seddog wnissnjog snouphyuy fo uoynossig fo way{—'][ A ITAVI,

PICKERING

300

19L9-| Z¢98—| gT#sT— 60r-< | 889-0 GLF0-€ €4- 119 91 | 9e2z
Zh6S—| FS88—| 96LFI— 098-2 | ZZ-0T 80- ST eg. T19 “OT | 882%
a % a I1SI—| 942-2 | L99-8T k6- 02 #8- 119 91 | 882z
O8FS—| GIO6—| 66FFT SSP1—| 96z-% | S2F9-8T 66-02 88-119 91 | 8822
L68h—| soze—| ZOTPT— GLSI-% | SEOF- 8z 199: 08 2g. T19 ST | S82z
SZIpI—| Z81-2 | Z9L-1z 6FG- && 96-119 St | 9822

s09r—| Fee6—| ZEGET— 1PF-1 | 92%-21 LOL-81 £9. 609 OU | 8k2z
e9FEI— 140-2 | Sh9-6T | (GZ-8T) SSTZ- 12 66- T19 : 8EL2Z

SiZh—| 9IS6—| SZLET—

ANSTO SROG

‘WwW “19 : s-ti ‘st “M ‘ejdmug

‘supib 629.9, = "jouw 08

€
'8L-0FF = O'H9"OS'H"OS"D ‘a7nyding soddop wmssnjoq poyosphywozy fo woynjossyg fo y2—TITA T19V],

%,
fo)
_—
m
<
7,
Loa)
ice)
=
o)
ro)
Hn
<
2)
ml
=
&
fd
=)
fy
fo)
B
=
fa
ee
-
a
fo)

(‘panurqUoo)—XJ ATAVT,

TLO- 22 (¢.6) %6%.2z €9-22 L0- #19 €0¢- OT 860e °
OLZ- &Z CRP. 2% L6- &Z | 80-19 ; LOVE

S61. &Z (¥6-6) 12b-€Z 9Z- FZ P19 . ZCOE

OFZ. &Z OSPF. &Z Z9- 12 P19 ° ZC0E

S96L-88 | (po. or) 2460- s€ 60-98 -809 F09E

GZS. ZE 1¥8- FE 1Z-¥E 2. 809 FO9E

9LI- 6F 89T- 0S #8- SP - 809 919-01 | 098

C89 -8F S01. 6F I¥- 62 - 809 GG. OL F098

GSSO.8F GLLO- 6F GZ -62 -809 8¢¢- OL $098 °
£20. $¢ (L1-#1) 999.99 ¥Z-t9 - 809 €19- OT F098
968-2 | (1-1) SISO. et 80: #F - 809 L&¢- OT FO9E
CLL. 98 128-98 90- 98 P19 G89- OT ZS0E *
SZ80. 0 606- 98 C166- 98 CL. SE r19 0S¢- OT GLLG *
Z9P-0 896.0F | (02-91) OZP. 1h LL- &Y - 809 T19- OL FOIE *
SI8F-0 GhSe- 8&8 | (ZF-91) 998.88 0-68 -809 ~LS-O1 FO9E *
161-0 S16. &F QOL. FF F- 1F -809 TI9- OT 709 *
OLT.O GZZ- OF C68- FF L8-&¥ - 809 16¢- OL FO9E *
140.0 6L9- ZF (48-81) OSL ZF ¥9- Le -809 T¥9- OT FOE *
OLI-0 ZI-22 | (g9.g7) 068-8 ¢9. 19 - 809 G09- OT FO9E
Z61-0 804-19 006-1 €L-2S -809 L19- OT FO9E °
610.0 689- LZ 0L9- LZ €e-1 P19 I8¥- OT T80€
L10-0 ZPE- LZ GE. LZ 1S-L% P19 129-01 ZS0e *
910-0 PL9- LZ 899. LZ 1-22 P19 €¢9- OT 9808
1z0.0 19-12 929. Lz €. SZ P19 0¢¢- OT CLLZ
6698-0 ZEL- 68 CZ98- 68 LE. 68 - 809 699- OT 709
ZrE- 0 916. 88 FEY. 88 €1- 98 - 809 €09- OT FO9E
C060. 0 CEPE- BE ZSZ- 8E 6L- 68 P19 109: OT CLlZ
CEs0- 0 C6LP- SE 968- 8 TP- 88 P19 899. OT GLLZ

wo
|
—
N

¥—)
nN

eNO Fis Ob
Sees ese

Ss
re
5
E
<a
(o=]
is
ay
=
fa
B
fe,
ro)
=
oO
A
iS
5
<
Fe
q
=

a ey ee ees ee ee er

raid SSR HAS

PICKERING

W Apt “A 2 | ‘wey, | “AA ‘ojdueg
103 — 4 | ‘ |
| |

= ‘18-TFT ='OS**eN ‘a7vydjng wnipog snosphyup fo uornpossig fo yve7—X] AAV],

302

x,
co)
ml
H
<4
vA
Lon
=|
—
2)
oO
|
=<
2)
—
=
a
jee
oO
x
o)
is)
&
ee)
isl
Z
o)

OT- &

TE. #2
‘OSTO: $B

€€09T—| 9ST6T—

PECIT—| SEI6I—

SPE9T—| LZET6T—
OOF9T—| 9ST6T—

LeI6I—

eL16I—
SvI6t—
SZI6l—

9806T — |CFEG.- S
9LI6L — |S2Z6-
O6I6T—| O88-¢
P6IGI—| $98-¢
PSI6I—| 9F6-¢
€6161—| $26-¢
SCI6I—| €86-¢
ZEI6T—| 006-¢
68061 —| ¢Z8.¢
T1261 — |S326-
T¥6-¢
126. S

ccese.- SE

06Z- £€

(4.08) CtO-8Z
CO6Z- 8Z
€1Tl-8Z

L¥G- FE

I61L-PE

6FE - FE

90€ - FE

OOP. LE

TI6I- 68

(94-02)
(IT- $2)
(20- $2)

sz. €29

} s0.¢29

0S. £29
1%- 29
€¢. 129
9F- 129
SZ -€29
L¥- €69
69-129
€S- €29

“L

“OL

O Bok

"MIOT I,

“M

LSO- ¥%

869. €%

2S0- FZ
806. £2
1ZL-¥2
9T8- &Z
668. &Z
180-12
L96.- &Z
$6. &Z

STIPE

SIFE ~

c60€
C6CE
|) SIPE
SIFE
c60€
c60€
STIPE
S608

AAG GON D

S
=

‘ojdureg

‘smvsb OOT.~Z = “jouw —

OF
€

‘Th-IZE = O"HOT°OS"*N ‘asvydjng wnipog payvsphyvoq fo woynpossig fo yoay]7—X A1AVJ,

LLOT—
stoI—

9TOT —

p6et—
7Sél—

TIOT—
Tc0L—
1s8—
£98 —

SCP.
Sc6s-
ZO9P-

L0Z-

eal.

€Z€.

PEL.

POL.

VOL.

ee ROONNNOOSO

ZP0- €
626. B
eST- TS
99P- bP
OSS. 8P
O8P- ZT
SOS- ZT
S08. 67
S19 -6P
OSP- OV

(eg. 2)
(g0- €)

(¢99. ¢)
(26-9)

(S6- ¢)

(28. 8)
(04. 8)

LOP-€

SZ. G

S19- €S
€L9 -9P
ZL9- OS
€08- ZT
828-61
6€¢. TS
648. 1S
PSL- SP

LOS
LOVE
Pogues
PO9os
PO9S
LOVE
Love
FO9E
FOYE
FO9E

88
“LE

“ce
“FE

‘TS
‘0€
“63

“4

BUSCA G

‘ojdureg

(‘penuiqU09)—XJT @TAV,

iS
ee
5
B
<j
ee
fs
:
ics
°
=
Z,
=
=
a
q
FA
E

7
.

PICKERING

304

80- 8T

STS. 81
‘O 1h-61

eesst—

€lI6t—

ssié6t—

O9T6T —
POT6I—

SZ16T —

L061 —

Z116I—
P2161 —
ZOI6I —
P9I6I—
io
86061 —
09061 —
92061 —
OLI6T—
8ZI61—
SSI61—
OO16T —
Z9I6T—
92161 —

€1Z61—
16161 —
26161 —
osi6i—
92161 —
8e16T —
6L061—
esi6t—
o9T6t—
P9T6I—
esit6t—
gsI6t—

OsI6T —
LLO6T —

$56-
S9L0-
€96-
090.
SEPA.
€L0-
€00-
¢80-
$900-
£80.
coc6-
SZ26-
996-
C8é6-
690:
$820-
P06.
cls.
TS6-
8&6-
SOT6-
926 -
Sa6-
O16-
£906 -

g
9
¢
9
g
9
9
9
9
9
g
$s
S
¢
9
9

S9€- ST
SILP- 9%
L8Z- €T
C98E- 9%
SO89- 92
SPE- 96
TZ8- 91
SZ0- 0€
SEt6- 9T
660- 0€
S¢00- 06
618-6
168-61
901-6
SIS€- 12
€9T- TT
Z9P- 1Z
2L%- 11
900 - €&Z
S618: ZT
886 -2Z
692-21
66S. SZ
068- ST
28. CZ
GLE. ST
Plé- 91

(98. ¢1)
(98- 81)
(68- $1)

008: 6T
8PS- ZE
0SZ- 61
SOPP- ZE
PZ9- SE
SIP: ZE
PL8- GZ
TIT: 98
0S6- GZ
S8I- Se
196- SZ
CPL ST
LG8- SZ
CPP9- ST
SEOP - Lz
SI6I- ZT
998- Lz
L¥I- LT
SLS6- 8Z
SLoL-8T
668- 8Z
£69- 8T
¥<S- IE
00€- 12
SS8EP- IE
992. 12
SO9E- ZZ
V6. FE
869- FZ

bez. £29

} 12: $29
©8- 829
8%- £29

bee. £29

} oz. £29

} 22-229

} 18- 829

} ev. £29

bor. £29

} 92-229

bor. £29

bor. 29
GL. 129
az &Z29

“ULLOTELT

“M

692 -

OST

68Z.-

¥%

“BZ

BLL.
9&6.

6S&-

STIPE

STIPE °

S60E °
C60E °

STIPE *
SIPE *
STIPE
SIFE *
SLPE
SFE
SIFE *
SIE
SIPS °

RIFE

S608 °
'SIFE “IT

‘ojduirg

LZ681— | 180-9 céL.9 18-9 €8 LIP-#Z | QTFE ‘OF
Cc —_— — >
SI5VI—| SEOST 8Z68T— | SET-9 oss.9 | eF-9 | ¢8 129 | 1¢9.4% |SIFe “Sy
c PERT — | 689. (8¢-6) 248-6 | 926 €8 079 | 208-22 |SIPE “HF
| se6st— L¥68T— | TTO- (zg. ¢) SO8T-ZT | F¢-ZI . LZZ- FZ \S1PE ‘ESP
9z681— O£68T— | ZF6- “ $90-@T | 66-11 : ZES-£Z | SIPS ‘ZV
} 98061 — | 0g0. (c18-9) SFSz-ST | 66-FT 121.4% |STES ‘IF
Leoest— 4 St68I— | FI6- (FS-G) GO88-IL | SF-IT ; 6Z8- £2 | e608
Z6881— | 2880. (g-G) 228-11 | 1-11 , 698. #Z \s608
ZEZ6I — |SSOT.- e6z1-9 — :
Soret — EFIGI— | 9ST-9 |¢z00- EI 891-61 | 28-21 $95.76 | SIPs
18z61— | 880-9 | 1F1-0- 116-9 — aioe
11Z6I— |SO9T-9 | F18- ZT CFL6-8T | 08. LT $67. Ve er
SSI6I— |¢990-9 | 6PP-T SSIs. £ —
98.8 sts 79061— | S11-9 | LOV-FT Z8S- 0% | 84-02 GOR. 96 |SI¥E
LE161— | 900-9 |E81¢- T Shs. L — |
S8I6I— |¢z60.9 |SFOS- F1 L6S-02 | $&-0Z 676: FE |SITE
*S-6 SLIG1— 06061 — | 900- peL-91 OFL- 2S | OL- 22 OI16- &z | e608
OFI6I— | 910-9 | 986- ST 200-8 | Sh-2z | 1Z-€Z9 | L10. $2 | S608
19061— |9000- 9 |S8Z2-% 642-01| — \ é
98.6 eenet< 9II6I— | 660-9 |S28z- LT G9L8- 8% | ST-€2 46-69 | GFE. 42 one
Z8I6I— | L&6- ZS3-P 68T- OT - } 0 ez9
OLI6I— | 100.9 |¢¢8z- LT G98Z- 8% | $Z-€z “G69 | 186. 8% | SIFE
a ZO06I — | 668.¢ |¢899-¢ yes-IT| — | \ RE
nadine — €0)61— | 926-¢ | OOL-8T 919-¥% | 19-92 | 60-€29 | 210-46 | SIFe
STIO6T— |498.¢ \cEese. 4 0122: €1 = } 7 Re
£0. II oco6I— (00681 —)| L06-¢ |GZ2zF- 0Z eres. 9% | 92-92 | 60-829 | 198. &@ Ste
: 9L06I— |SEE0.9 | 8Zz-Z S198-81 | = bus exe | ese.0n |
S1I6I— \cFZt-9 | OVS-0Z crse- 92 | 18-92 "869 | SLP- 0G | SIFE
; ae oe S6161— | OTO. iP. CceP. OT — 91 } - ’
—" craton 1u061 — | 820. G- & €1-68 | 64-8 eg | f 0 £29 | OOT. Fe |SITE

DWrM wD MMO MW OOOOOODODOOOOOOWVW OW OW

Zz
S
_
E
=
q
2)
=
=)
s)
4
-
2
=
fi
ea!
rs)
&
°
H
2
<>}
m
&
a
=
z
)

‘L W | “OT . . 4 “ “L “UMeT] “MA " ‘ojdaieg

XUIX.

VOL.

(‘penurjuvd)— XX AV,

Q
re
5
=
ys
fa
Ay
ra
i
B
fy
a)
2
oOo
Z,
5
|
fe
q
&
joo)
=

PICKERING

306

‘aoquiry aoddn oy} OUT poaomter useq Sutary Axndso9ul OY} JO OUIOg »

C98. 2

0sT9—

€ST9 —
e919— 9219

v8SI9—
1619 —
2809 —
té19—
e919 —

—
$t—

| OST9
| 6819

¢so- 1
1Z48-T

86L- 62
£98. OF
Cg0-T | 882-62
Ce6-1T | ZbL-08
ULS-ZT| ZtP- 68
CZ16-Z1| O18- 6E
SIF9-ZL | FEZ- 6E
9¢0-T | 8LZl-18
C668-1 SPSS. &Z
OL0-T | 8L4-&2
uss: T O9L- &Z
180-T | @¢2-18
yS8-T | 682- FE
0-1 | 1hL-18
G18-T | 192-8
980-1 | 286-28
928-1 |S00S. 98
€hO-T | SL6-ZE
0¢8-T |STLP- 98
GCIS-Z1| £16.68
CCFI-Z1I| 1Zb- OF
O&Z- Z| S88- 68
8Z0-1T | L1S-€8
€88-1 | 92h. F8
810-T | 62-82
SI8-T | LII-62

€cs8-
cel.
sI8-
LL9-
Z10-

(18. 22)

(06-22) SZzZ-
Ccs8

628-
PCP:
8P8.-

OT9-

684-
SPI-
F8L-
cri.
€20-

(#0. $2)

S1cE-

Salé-

99S
cgs.
crs.

(LZ. 2)
(gg. $2)
(ZF -$2)

€18-
Sé6

coZE -
810-

606 -

0€
a

6S
es
rE
6€
62
0€

06- ZE
€6-ZE
| OF. ZS
| ZP-SS
~9-ZS

¢8. 3
18-92

08- 9€
¢9- 98
¥ €. 88
¥e- 88
PL-ZS

99 -2S
¥6- 6S

SF. 6E
LL-62

6
6
68
+8
08
#8
#08

|} +0. o19

} ¥0- O19

OF- 809
OF- 809
OF- 809

} 60. O19
} 60. oT9
} 60.019
} 60. O19
} 20. O19

} 20. o19

cP .809
CP. 809
CF. 809

} 80. O19
bor O19

€L8-Z1 | go9E °

seer €09E *
6F0-€1T | &19E *

StP-EL S8ZrEe

$80- €1 | S09E *

€Z1- $1 | £098
ae a
sn ielain
P6I- ST | S098

4Z0- €1' $098

Z81-€1L; S09E *

9ZI- 1 | SZEE *
SZFE °
Z$- 81 | €198 °

80T-€T | £098 *

0S6- ZT

Notoor Dd &

260: &T | 09 *

*payoo100 JT

“WOH
00110)

WwW

ore “A

"mAOY J,

“M

“"

|
.
|

‘sw1b 10.81 =

‘jou —

OV
&

‘88-8241 =

'OS'H ‘apy djng wnssnjog fo uornnjossig fo ma]—]X «avy,

%,
=}
—
e
<
Zz
—_
ioe)
A
°
oO
4
<a
2)
=
S)
&
fo)
B
a
<2)
jem
13)
es
il
Z
i)

£68- FI |€€89—

P-ST |26L9—

ST6-9T ™ aa

Té-06 cEeeg—

|
‘00T.12 \4929-
|

9889 —

0289 —
roto {
0289 —
299-4
19s9- {

(si¢e9—

PrEesg—

Tet9—

6Lt9— {

6849 —
6189 —
OF89 —

9189 —
11Z9—
$189 —
L9L9—
PPI9 —
2999 —
Z9S9 —
19s9—

8Lt9—
£09 —
£689 —
6hF9 —
¥0¢9 —
68F9 —
F8t9—
9FEY—
cZe9 —
#829 —
0¢Z9 —

8t—-
8se—-
si +

8t—-
8t—-

2899 —
0049 —
0099 —
66°9 —

9Ts9—

ZS9—
LL¥9 —
Y9F9—

ZZE9 —
8829 —

260-3
TI8I-@
SZ9T-%
OF8- ST
TSI- T
210.2
F6I-T
P60- 3
¥C0-%
£60: @
€20-3
S080 -%
£62. €T
890-2
986-T
Z01-T
€Z1-T
c9L0-Z
S610-%
LS0-%
ccc0-.@
OIl-T
T46-T
CPG. T
$266. T

COP: OT
C198- 02
CLZL-0Z

OSP- 8E

880. 6T

€8Z-11

980-61

c69- IT

9T6- FL

ZS. GS

9Z8- OT

€00- LZ
Cs9Z4-8€

980- 12

6S0- 12

SIE- FZ

L8Z-¥Z

908- TE

696-02

68Z- 1E
cccl. 12

¥60- 92

CES: SZ

LZZ- 92
C961. LE

LGV: BI

(G8. #1) SZO- 21
O68: ZZ
062- ZS
69% -0Z
GS8. S18
0€Z- 0Z
682: ST
0F6- 9T
C89. LZ
6hS- ST
GE80-6Z
GE90-ZS
PSI. &Z
GhO- &Z
OZP- SZ
OIF: &%
CZ88- EE
v1) S8L6+ 2B
(S10-61) OF8. £2
(90-61) II8-€z
$02 -L2

€1Z- 92

G69T- 8%

681-68

(¢0- ST)

(¥9- 91)

(6T- L1)

(9-81)
(8. 81)

(6-81)

} ee. 609

06- L09
8Z- 809

} 26.609
} 26. 609
bis. 609

} Z8- 609

€€- 809
66-09

} 28: 609
} 28- 609

bes. 609
66. 209
} 00. o19

i 88. 609

CIO- €T

806- ZT
S°6- ZT

600. €T
8£0- €T
164: 21

606-31

120-&T
Le1-€1

026. ZT
PST- ST

$00 -€T
910. &T
ET - I

916-61

oO
GELZ IS &
cZlZ ‘OF
COVE “62

€09€

S8ZPE
Gols *
€09E °

SZrE “LT

*pawor10d

uot}
-00L100)

cad

“4

BUACIIAg

“mM

‘ojdueg

*(pannyuod)—TX AAV,

€c6L— Ch C6S¢.Z ‘cont. IT | (2-2) 904-8 . T8- 09 | 266-21 | SZzz
Z16L— 68L—|O019S-% |S629-% | (10-€) ¢0ZZ.¢ | €9. 18-209 | 890-81 | ¢zZz
1164— O0E-9T |GhFO- 98 | (66-8) CPPE- ZS | FF Te 11-809 | 9F0- 1 | S19
_ J e99z—- : C8t-Z |CLZ1-6 CZ19- IT :; €8-L09 | S66-Z1 | 8ZFE
6994 inn e— | se9z—|voer-z legoo.6 | 82) “ces. 11 | ee. 8-209 | €z8-21| szve
{ GOSF-Z |¢go0- CGF. c. Gg . LOS : ZVE
2691 — O0&L-ST| 80L-Se \(C96¢%) SEF-T¢ - Of 1-809 | ISO. €1| go9E
St9L— l 9S2-S1| 289.98 | (Sh-9) LEP. 2g 2g SI-809 | 8F0- 1 | S19E °
{ozes— { 9884- 9¢ SoF-Z | 89Z-ZI (g9.9) 984-F1 | BI-¢ £8. L09 | 880. SI | ¢zZzz
. LysL— aS CZLF-Z \S81S- ZT 16L-F1 : L8-L09 | S61. &T | Szzz
— CEI-ST| OLL-2E £06. ZS | 20-2 81-809 | £86.21 | go9e
c0Zl— C8PE-Z |CE98-0z Z1Z- &Z | 90. $Z 16-209 | €91- 81 | zzz
19@L— -L—| OF€-% | $69-02 FE0-8Z | CO- Sz 16-409 | 10-81 | ¢zzz
csIZ— Cal, ZhE-Z | SLL-2z CIS. FZ | LE. Z Z6- L09 | 290-1 | SzFE
LSIL- ; 818-2 | L61-2z (eg. or) STS: 6 | 8-42 Z6- L09 | 816-21 | 8zFE
181Z— L 9FE.%Z | SHI-Zz | ** 16%-#% | Zh-Z Z6- L09 | 260-81 | 8ZFE
0sts—- G1ZL— _ Zz Lee-Z | 61-22 08S. FZ | FE. ¥Z 26-409 | 86-21| SZFEe
0¢-FL| 028-2 |(eSF-01) O88. zg 2S 02-809 | 688- ZT | g09E
c1e-% | 290-82 | (GZ-01) LLE-¢z -¢ 06-409 | LOT- $1 | zzz
€20L— — I¥E-F1| SST- 88 9LP-ZS | Sg. $Z-809 | SEO. ST | £098
62-2 |zzZ- 18 C186- && . 68-209 | 8ST-€1 | czZz
( ¢969— 96 L8Z-B |SSTL-1€ |(oog. 97) S6L6- 8E | L9-¢ -L09 | z¢0. €1 | Szzz
; 092-2 |GLZ8- 18 |‘“""" °’ SLEI. $8 ao 3 68-209 | €1Z- $1 | ¢zZz
‘O96. S1\ZS69 — ; 682-% | 989-18 GZ8- EE | 98.FE 68- 409 | 90T- I | £643 ‘YS
, ¢9Z-$1| 28-2 | (96-81) 880-29 | 8Z-z¢ 92-809 |T01-81 | Gzzz “ee
Z9E- FI |SF99- LE €92Z0-2¢ | &6.6F 9Z- 809 | 091-€1/|SzFEe “FE
OLZ- FI \Sze0. 8E | (S6- ET) Seok. zd 2 92-809 | &I- ST ee "ee

{<3}
om
=)
a
<
a=}
=
re
13)
a)
hy
o)
nm
oO
is
uw
=
Z
mH
153)
ee
Sa)

PICKERING

“u0r
“paqoa1100 mus W ‘4—3 “4 ‘mu9uy; =A “"

‘ajduing

308

(‘pan U14U09)—TX WAV,

Zi
S
—_
=
<|
AZ
—_
ze
a
3
oO
1)
_
ral
is
q
oO
fis
3°
4
=
2
x
a)
Z
3

€P- €T

89. ZT
L4Z-81T

9-12 |
|
|

"098; 22 |

LT19

€629
€S¢9
PLS9
T¥S9

698-2
L9¥- 81
PSP. SI
TS. €T
ZSZ-
ZPE- 2S
(c8- 1) Ze. 88
(ZO. 81) 949- Te
022: 02
ZSL. 1%
LLY-08

889- €€

Z8-T

22- ST
¥6- ST
61- FT
09- $Z
88-12
T9. €8
60. ZB
62-02
98-22
~S- 08

26. GE

66- €19
00-F1T9
09-19
00-F19
SO. F19
¢$0-V19
II-¥19
It- 19
02-19
1Z-¥19
92-19

LZ -4%19

|
|

8Zé-8
02-8
¥LZ-8
L61-8
08-8
chI-8
9FT-8
$66.8
c9T-8
0€2-8
968. 8
02-8

x © O

a ao # ww Ss

90FE
9808
90F€
£608

90F8
PLOE

"u0Ty

7 JIILOS
poy WwW | -OOLL0F)

*L | BUCCling | “MA

‘oidureg

‘swn1b 897.3 = “Jou —

OV
€

‘

$8-601 = "OST ‘apydjng wmywvT snosphyup fo uoynpossig fo yoa}7—] TX ATAVI,

| :

ee zest | 1162 cezt-t | 9g¢-ze | Sz9r-1e | 28-ze £8 99-019 | 926-0L |**** eeee “ZI
SD o8-et | 989 | SPst | sels 0-1 | 96n9-6¢ | S619-e¢ | 6F- Ee eg | ¥e-809 | 804-6 |*'** Slze ‘9T
4 ‘O° HOS tI
el
& a . . - r _
E |**ezog pus
<= 8-2e | sozet | zt9t | t160zt — _ ‘QOL ‘eget “suvty, “f 90g ‘mayg 208 spp 20, 4 | ‘ZE0E ‘BSLZ
a
fs 61% ‘SI—L
= ‘O° HOS?
i |
fod ”
: :
~ gig | oes-2 | eze-pe | pep-1e | $9.58 eg | 68-ST9 | ogF-et |****¥eLz ‘9

00-6 | AvGOT | LEST | OTLB) Toye | coz-z | pee-pe | osc-1e | 98-48 €8 1e.¢19 | 2ez-et |**** $eZe °¢
8 ecss | 6Is-2 | ee¢-ez | P1Z-02 | $4-6T 18 lp-F19 | 188-8 |°°°' POLE F
z SAT | OlFOT | UST | B68 1 cgce | ste-z | orv-ec | s29-0z | 92-6t | 18 | s-FI9 | GLz-ET |***"FeLe "8
ts { #006 zos-2 | 169-98 | 6eZ-e8 | FL-¥8 18 1Z-¥19 | 82P-8L |°°°* PELE Z
ry «0 O886S | ~TILOT | B29T | 6806) ciig | geg.z | 129-98 | 682-88 | FL-F8 18 12-619 | 9o8-eL |*'** ELE ‘I
= ‘
& ‘O° H*08"90
: a
& ‘L ‘Ww | * “ut 4-4 4 | 3 | “4 | mou, Mm || copdunng
: |
= “ $896 = . 08-261 = O°H"0S"1
ta 6c8-0T = 22-L8l = O°H*OS4N
Ss) OF ‘
Ay ‘swvsb 09Z-ET = Jou — 608.921 = O*H*O0S8"0

€
“mnryzvyT puv ‘wnisaubvyy ‘saddog fo sayvydjng paywsphyouoyy ay; fo uornjossiy fo yooyT—'[] IX WAV,

310

Journ. Chem.Soc. May I886. Degrees Centigrade Plate 1.
1 3 4 7 9 n 3 15 17 19 21 23 25

20,800 20.800
J
>
20,600 Pe 7,800
——o wees wee P —_
20,400 A 2_|7.600
_ HEAT OF DISSQLUTION OF FHE MACNESIUM SULPHAITES |
° :
| j ae a Cae WA
20,200 | wo 7,400
| | | | td
a on eee 2 oa an aes tee 4
= te — a — o | — Pa _ i ie aes 2 a a a a
= 2 TC F A | | @f | |
* en : a ae eee: es a ‘
20,000 | gf | | __|7,200
| | |
4 T - _ -4 ad
' 4 4—-—-—-4- - 4 —-——_}+— : —_ + ——+ -— — i» ————T — — 4 _ - ad
ZZ re A ‘a [em em ‘eon S,
79,800 L 7 7,000 ty
Ys }

+ r “nt A T 4 + 4 ~~
a es oe oe —}- —- —- 2 ~
is -—_—_+- -——+— + ~ + —— - ——4-—— -—4-- — —_ 4 + -- — 4 ‘
19,600 [ os | Tt = 6s0 8

| 7 td ws . =
§ = oe a “a — — cox mane nee mah cane [ tie 3 —+——+—+ 3
oa os a
= Z FZ i SS a AS NE - :
2) : a)
& 19,400 | Z 2 6,600 5
8 mer 4 iS, | oll +
S 7 | S
= iow -7f adheres -+——- —- — + —— 4 3
B + 7 + 7 = 2
RY 19.200 Z 4 6,400
7 om . 5
LZ 2 A SH | 5
-——— — VA —E 72 ——E — $$$ — 4 -___4}—___} 4 .
19,000 vA 3800 6,200
+ —t— - A eens > =
a: a A a
4 1B, |
Pa a 2
19,800 L ales a. 6,000
2 te L dasdolule™’ = + —t-—}—- =
——- dines an ane ‘= ——
= — ——+—--——+——__+——_ --4—-_ —-+ = - ————}F
l 00 | 5,800
13,400 13,400
= + ‘ +--+ — = 4
= a a
a ._ —— = 2A Ee e ——————
13,200 = eee ae | 13,200
= ———EE Eee t—
—— HEAT OF COMBINATION OF U,SQo4 H,0(sptid) =
| | | | ;
13,000 | | L L i 13,000
1 3 5 oe 9 1 13 15 17 19 21 23 25

Harrison % Sons, Lith. S* Martins Lane.W.C Degrees Centigrade

Plate 2.

OHS 09N9 Jo uogmqossyp Jo rosy anaz

71,000

+ —-——++

—f- — —+4—-——- 4 200

$$$ —_—__j _____f

—+ +--+

——j—_}_} |}

—

10,800

Se Ee

++ —

1 a Ae ee

4 ,
mn
a

OMBINATION OF Cus

wr.

Pe
Lo

HEAT OF ¢

Degrees Centigrade

I
T
|

—+——$—$—$— 4} —____

|
j

—_

| HEAT OF DISSOLUTION OF [THE GOPPER SULPHATES

—+—-++—

——___—_—_+___}

16.6002

16,400
15,600)

“
S°
=
g
“
&
S
S
3
S

15,400

08 9 Jo uorMossrp 40 PoaTy

15,200
11,200

14,400

11,000;

.
_|

|
|
13

Degrees Centigrade

‘ Martins Lane,W.C

Lith,

Yons
20ns,

mison *& ¢

Tay

OHS ('08)UW $0 uqrossrp jo yoay anay

= >
Ss 3 &

= ; &

t—-

a ke Sa

\
i»
3
;
|
|

ss)

SULPHATES.
OG ae anata,

sium

HEAT OF COMBINATION OF|M,K.($0.)$6H. 0(salza)

HEAT OF DISSOLUTION OF THE POTASSIUM MAGNE

S S
= =

= sy *

Sar so uonmpossrp so qoayy

n & Sons. Lith. St Martins Lane. W.C

Harrison

Journ: Chem.Soc. May 1886.

Journ. Chem.Sec. May 1886.
- 7 3

10,200

|
|

———___—_ —_4-_ __4— |
AT OF DISSOLUTION OF UM-COPPE

—+___4___}__ —

|
|
|
i

+
|
|

Seedinedionieed

True heat of dissolution of Cw K.{80,),6H,0

=
y
=
2
:
H
3
S
:

—_——_+—__- —

oe
F POMGINATIDN OF LUl

13,400
1

Marrison & Sons, Lith. St Martins Lane.W.C

Journ Chem.Soc.

o*
D.
v
=
S
$
3S
‘Ss
%
%
a
>
=)
»
3 ;
ss

‘

UTION OF THE SO

1
+400
ae mee
+200 |
0

Degrees Centigrade
13 1
| | T
ae - ; |
| | |
== ee ee a + ae
a oe oe oe oe 4
DIUM SULPHATES |
" ;
+ —1
|
~t—
-+- 4
San
|
1
ae
4
. 4

=] 2,600

%,800

3,000

3,200

3,400

_-|3,600

4
\-IR9N0

—e

+12100_|

3,200

4,000

4,200

3,200

oe TrTy
!

3,000

|
[|

N OF we SOHOH (solid |

3,000

True heat of dissotution of Na,$0,i0H,0

Plated.

Fay

"15

ae
tx

=? 7
sod

~
-

2 6.

Degrees Centigrade

Sourn. Chem.Soc. May 1856.

039°Y 40 uoynpossxp jo yoary

8 5 5 3
S > S
= 5. = a = S § : % 8
—— “S — 6 Cs) = SS S nS SS ~S oS
1 | | ' | |! ' ae | TT] TTTTIVTTTIgT tI ) i
| | j | | | i |
a a | \ : 4 ; |
oe } | } } |
—+ T T --4 + + p—+--++4+-4 —+ a 4 +——
| ! | | { | } } } |
i | | i | | |
rT rH CTT ttt)
1 | | | | PET yy Ey ty dy]
ecdaad | = | | |
tT + +4 t +— —+$——+ - aofinneadionnil — + ———T
| | | ; | | |
ae | imo |_| Lt jf} | |
aR | rT mi 8
| , | | | |
| | | |
a & = : a = oe ee | } Daag i
SGGSGGRRERGRRGERGRAGn
| | | | | Pi} dy yd | | |
++ Lesh } F |
, + T T | T ] = | i 8
| | | | ris d dd Td dy
es . £ 5 ; | ; | | ; F
=o +++ + } 4+ +--+ ——-+--++-- =
LN WN ae | PELE
: Lid | ann Li | |
| Y rrr BSUSSERES EERE &
} | } | | | | |
| | N i | | | | os. o
“- TI "Ty TTT Tt ine TT | TTT
| | | | J | | | i ; | |
= i ee aS jo} RBar”
. i ne mi TY] rT. so
| } | | ; | ; 4 i, a i
i | ; i i
iP | | | | | | | | 1 | } | } }
rs ——— =e ei | +++} |} |} = se a = | +
fan) | | | | } | | | | i | } | | | | |
9 ears | || | | Ltt | | SaaS ae8 2
4 1 | i | 4 4 4 Cr)
i i ~ T | T 7
2 ry i [| IN TTT | Banal rT {| PT TT yt Tey
| |e . | | | | | Y | | ae |
ne 6a zs t 5 Son an Bee ee ee
| } | | | j ! |
E | rT TTT TTT PPT TTT TTT ty yds
i= | me } | | | i 4 | | | | ®&
rst + TT oe T t TI a + T 2 »
3 | } | | .. 9 | im , 4 } 5.
5 | i | 1 i il Ltt tit =
rary — +++ ptt ttt tt
| ze | | Pd ; |
| | | | | } | |
: ae NI eee! Reeser
| To : ime ame pony - a oe TTT 15
. ; | a | ; | | } | |
| 1s } i | | | : t. |
| | | | | =
+ 1 : = oe fon ee Oe
| Te | . s |
| |g | BAae
it = | bk. a
a. | } | |
a | | | }
> | | | }
1491 . a = a
Py Tw rr BE.
| |x| | |
Lee | |
ik TT
| Id | |
| lz
++ Oo + ~~ =
l-
| |3
| i j }
| J | t
| IW” | | j
| |2 | | |
oo S88 _|
| re] i + ——+— 4 —
S488 | | |
2) eee : NS
} | tad | | | } | }
| |x) } | | | | | N 5
| ; | | | | ; | } | i
BS EER RO EE REESE EGER GEESEEEEGEREI :
| } | 1 | | | | | ”
| | | | | | | | 4
S = i i | 1 iL | j he i 1 4 4 i ”
> > = > > S > S ~
S
= S S 5 S S : 5 S S = :
te © © © © © So ry “ “> 6 iF
> 2
OS 'TJo UoRNOSSrp Jo YoaryT -

Plate 7.

ay 40 T40.1b 96-1} Jo UOISNY Jo Yory WaOT

g 8 8 & : 3 S S e § E
nw ~ ~ ~ ~ ~ ~ ~ ~ = ~ ~ ~M
eR T ] = } T | | “7 | ] T Z if R

| | | | | | |
| " “7 a fo —r dood 4} TT —
}

|
—

—

= —

| |

|
—_—_}+—_—+

| coal ;
P

[—

atoboa se Ge eccabed

44 ___}

21
=
+—

|
Sm
=
, —_-_f/-- _
+
'
+
|
+
a
1
|
|
—
|
t 4
|
—
|
- &
—-
V4
4
}
—

i“ >
12,000 3
SS
Sos
TE 3a
CD
~~
|e
Zan

~
|
+

71,800 | >

12,21

|
|
4.
.—
CS
|
i
VA

|
i

14,000

- s S
Phony OTH” Pritt |
Po Hobo he , EN aeeee

|c 73,800

759,
Pa

|
4

LeHo

On Th

mast
'

Degrees Centigrade
15
SULPHATES AND LATENT HEAT OF FUSION OF ICE AT ¥°C.
ff)
——
S
SE
—
1
|
4
a
T

1 3
S
i
¢ N —- enue —— 7 -+—
2 N 8
, ! NS ‘
BAIL LLL LL oy Lf NN >
‘ + t 7 7 +— T < 5 + 4 |
| | aN
™- s 1 a ~
jit § L § \ im
wl H
= 18 =\( 7 % \ .
o vo
ar }_J e ! \ ke
+z — _ +— — oe © 7 =a — + ; i ee am
: LIT : ALT DAN TY
is 3 t = 1 53 | 1 i Ss
S | ! +.\ b | |
Ln +——_-+—_+4— N | ienaaiiaatdl = + 44 +--+ -+ | + —- dia = —t—-+ ii —+ _| +——
& NS | r ! N\ | | |
° | | \ Na | ,
“Ls aX SeHeRan CNH?
° | N\] &
eh LANCE EEE EEN EEN,
=| | | ) | LL LEN E
o + | |} | j =
| | a : r | | | | ry} 3
L | joo}. 1% | || | NX 5
| | ‘ | | ; | | ; | 3
— ze t Lid g s j | | il 2 | | LJ | 3 | & WS | | | | “~ 2
2 8 S 8 = $ S
ys Se S g , ’ A
‘os*u<0S Ws0 uommunquioa jo WoaR ‘Os wOS Ng 40 UoTGMNGuoD Jo Poa

Journ. Chem.Soc.May 1556.

SENIER: CYANURIC CHLORIDE AND CYANURIC AOID. 311
Taste XIV.—Heat developed on Mixing Solutions of Magnesium Sul-
phate or Copper Sulphate with a Solution of Potassium Sulphate.

Strength of each solution = 1 molecule to 400 H,O.
Quantity taken 307 c.c. of each.
Thermometers employed, 81 and 83.

A.—Magnesium and Potassium Sulphates Mized.

Observed heat Heat developed in | Corrected heat Mean

development. blank experiments. | development. | temperature.
YT} + 0 cal. + 63cal.| — 23 cal. 7°07 ©.
ee i } +5, |tpew bie, |] -8, 13°32 ,
a » bai me le » baw ae + Be 23°0 ,,

B.—Copper and Potassium Sulphates mixed.

1. + 104 cal. + 102 cal. . .
OF ys }+ Meal. |7 199%" f +107cal. + deal. | 10-6°C.
3. + 2 ,, a :
oe » he 36 ,, + 38, 2, 13°22,,
5. — 19 ” — 16 ” 4 | °
ee » b- 13 ,, = » b— 35 »| +23 5 22-89 ,,

XXX.—Contributions to the History of Cyanuric Chloride and
Cyanuric Acid.

By Aurrep Senter, M.D.

Recent experiments on cyanuric acid and melamine have shown
beyond much doubt that these substances are normal, and not iso in
their constitution. Nevertheless, it seemed interesting to extend the
data upon which this view rests by the preparation and study of other
cyanuric derivatives. With this object, I have investigated the action
of cyanuric chloride on the salts of various organic acids.

Sodium Acetate.—Ii this salt be heated with cyanuric chloride in a
closed tube for eight hours at 100°, a reaction takes place, and on
opening the tube the characteristic odour of acetyl chloride is per-
ceived. By treating the contents of the tube with cold ether, the

312 SENIER: CONTRIBUTIONS TO THE HISTORY OF

acetyl chloride is readily dissolved, whilst any unchanged cyanuric
chloride, being but sparingly soluble, remains for the greater part
with the residue. By repeated washing with hot ether, the whole of
the cyanuric chloride can be removed and a yellow residue is left,
mostly sodium cyanurate. This when dissolved in water gives, on
the addition of hydrochloric acid, a crystalline precipitate, which
readily responds to the cuprammonium test for cyanuric acid.

The reaction, therefore, is quite different from what was expected.
It was hoped that a triacetyl cyanurate would be formed thus :—

C,N,Cl, + 8C,H,0°ONa = C;N;(O°C.H;0)s + 3NaCl,
instead of which the following took place :—
C,N,Cl, + 3C,H,0-ONa = C,N;(ONa); + 3C.H,;,OCI.

By converting the acetyl chloride into acetic acid, and finally into
silver acetate, the proportion formed was ascertained to be 22°5 per
cent. of the theoretical yield.

Between silver acetate, in place of sodium acetate, and cyanuric
chloride, no reaction was observed, and in the same manner cyanuric
acid itself could not be induced to act on acetic anhydride.

Sodium Formate.—This salt behaves in a manner similar to the
acetate. On opening the tube a gas is evolved, which on examination
was found to be carbon monoxide, whilst the residue contains cyanuric
acid in abundance. Apparently formyl! chloride is produced in the
first instance, immediately, however, breaking down into carbon
monoxide and hydrochloric acid. The reaction is therefore —

C,N;Cl, + 83CHO-ONa = C,N,(ONa), + 3CO + 8HCI.

Sodium Benzoate.—In the benzoyl series, the same general reaction
takes place, but whereas in the previous instances it was only partial,
here 88 per cent. of the theoretical yield of benzoyl chloride is
obtained. The reaction is expressed by the following equation :—

C;N;Cl; + 3C,;H,O-ONa = C;N;(ONa), + 3C,H,;OCI.

Action of Cyanuric Chloride on Benzamide.—In the hope of obtaining
a benzoyl substituted melamine, cyanuric chloride and benzamide were
heated in a closed tube for eight to ten hours at 100°. It was thought
possible that instead of the desired action melamine and benzoyi
chloride might be formed. But on opening the tube, there was a
distinct odour of bitter almonds, and treatment with cold ether
yielded abundance of an oily liquid having the character of benzo-
nitrile, and easily converted into benzamide by treatment with con-
eentrated sulphuric acid. The residue in the tube was almost pure
cyanuric acid. The following action, therefore, had taken place :—

C;N,Cl, + 3C,H,'CONH, = C;N,(OH)s + 3C,H;'CN od 3HCl.

CYANURIC CHLORIDE AND CYANURIC ACID. 313

Notably, the cyanuric chloride, acting like phosphorus penta-
chloride, takes away water from the amide, leaving the nitrile, while it
is itself resolved by the liberated water into cyanuric acid and
hydrogen chloride.

Action of Benzoyl Chloride on Silver Cyanurate—While I was
engaged in these experiments, J. Ponomarew’s “ Zur Frage iiber die
Constitution der Cyanursaure” appeared (Ber., 18, 3261), in which
the author describes triacetyl cyanurate, one of the substances which
I was seeking. It is obtained by the action of the acid chloride cn
silver cyanurate. By this elegant method, it seemed easy to prepare
the other acid derivatives, and, in fact, it has already given me the
corresponding tribenzoy! cyanurate.

Benzoyl chloride has no action on silver cyanurate in the cold, but
by heating a mixture of the two in theoretical proportions, in closed
tubes, for eight to ten hours at 100°, an action takes place, and on
opening the tube much of the odour of the benzoyl chloride is found
to have disappeared. The contents are extracted with boiling
chloroform, which on evaporation gives a crystalline precipitate repre-
senting 25 to 30 per cent. of the theoretical yield. This recrystal-
lised from chloroform, gives colourless needles of tribenzoyl cyanurate,

C,N;(OC;H;0); — C.,H5N30¢, which requires—

Experiment.

The new compound shows no distinct melting point, being decom-
posed on heating. It is but sparingly soluble in chloroform, and
insoluble in ether. By heating with water, it is converted into
cyanuric and benzoic acids, this being effected very readily if the sub-
stance is enclosed in a tube with water and heated for a short time at
100°. This behaviour with water, which Ponomarew observed to
occur in an analogous manner in the case of the acetyl-derivative,
characterises this class of compound as mixed anhydrides.

In conclusion, I may mention that two experiments were instituted
to obtain two cyanuric nuclei joined to each other. Cyanuric chloride
was brought together, in closed tubes, with melamine, and also with
silver cyanurate, but no reaction took place at temperatures from 10°
to 250°.

University Laboratory,
Berlin, February, 1886.

314

XXXI.—Contributions to a Knowledge of Cyanuric Derivatives.

By Harotp H. Fries.

By the action of ammonia on cyanuric chloride, Liebig (Annalen, 10,
45) obtained chlorocyanamide, C;H,N;Cl. The corresponding aniline
compound was discovered soon afterwards by Laurent (Ann. Chim.
Phys. [3], 22, 97), and quite recently similar chlorine-derivatives
have been obtained by various investigators, especially by A. W.
Hofmann (Ber., 18, 2755) and Claésson (Bihang till K. Svenska Veet.
Akad. Handlingar, Bd. 10, No. 6), who acted on cyanuric chloride with
amines. The final action of ammonia, or of the amines, on cyanuric
chloride, is the production of simple or substituted melamines. Thus
it is clear, as foreseen by the above-mentioned observers, that the
compounds still retaining chlorine are intermediate derivatives between
the trichloride on the one hand, and the melamine on the other. Take,
for example, the methylamine series of substitutions :—

(C:N),Cl. (C:N),(NHCH,)Cl, (C:N),(NHCH,),Cl.

Cyanuric chloride. Intermediate compounds.

(C: N),(NHCH,)s.

Trimethylmelamine.

All of these are known, with the exception of the first intermediate
compound ; this dichloro-derivative is also wanting in the other series
which have hitherto been studied. The experiments with «-naphthyl-
amine, which form the subject of this communication, have, however,
given all these compounds. Similar results are to be expected in the
case of 8-naphthylamine, with the investigation of which I am at
present engaged.

Primary a-Naphthylamido-cyanuric Chloride.—Solutions of cyanuric
chloride and a-naphthylamine in dry ether are quickly mixed together
in the proportion of 1 mol. of the chloride to 2 mols. of the a-naphthyl-
amine.

A precipitate of a-naphthylamine hydrochloride occurs, which is
removed by filtration, and the filtrate is evaporated to dryness. The
residue crystallised several times from hot alcohol gives colourless
needles of primary a-naphthylamido-cyanuric chloride,

(C : N)3(NH-C,.H;)Cl, — C,;H,N,Cl.,

melting at 149°, The formula requires—

FRIES: CYANURIC DERIVATIVES.

Theory. Experiment.

ages on
156 53°61 53°47 — a
8 2°75 3°01 _- -—
56 19°24 -— 19°05 1894 —
71 24°40 ~- - - 24°44

oe ——

291 100°00

Secondary a-Naphthylamido-cyanuric Chloride.—In this instance, the
ethereal solutions are mixed slowly, drop by drop, and in the propor-
tion of 4 mols. of 2-naphthylamine to 1 mol. of cyanuric chloride. The
precipitated a-naphthylamine hydrochloride is removed as before, and
the residue, after evaporation, recrystallised several times from hot
alcohol. In this way, colourless needles of secondary a-naphthylamido-
cyanuric chloride, (C : N)3(NHCyH;)2Cl = C.3;HisN;Cl, are obtained,
melting at 215°, This requires—

Theory. Experiment.

| see cr —_
Crs cere 276 69°43 69°32 —
Babess 16 4°02 425 —
Mibaswes 70 17°60 — 1751

Cl.... 355 8°95 —

397°5 100°00

The secondary compound is much more soluble in alcohol than the
primary derivative, and by taking advantage of this fact they may be
separated if they occur mixed together.

Tertiary a-Naphthylmelamine.—Tertiary a-naphthylmelamine is
formed when either of the above compounds is heated for several
hours at 100° in a closed tube, together with the theoretical propor-
tion of a-naphthylamine. The melamine found is washed with alcohol
and water, and recrystallised from chloroform. It melts at 223°, and
has the formula (C : N);(NHC »H;); = Css3Hy~Ns. This requires the

following values :—

Experiment.

ig Y

18:33 — —_
4°76 490 — —_

16°67 — 1667 1685

504 100°00

Thus, in the case of «-naphthylamine, the whole series of deriva-
tives theoretically possible are obtained, and there can be but little

316 WATTS: THE ESSENTIAL OIL OF LIME LEAVES,

doubt that the primary amido-compound can be prepared in the case
of other amines, and with ammonia itself by working under suitable
conditions.

University Laboratory,
Berlin.

XXXII.—On the Essential Oil of Lime Leaves (Citrus Limetta).
Preliminary Notice.

By Francis Watts, F.C.S., Chemist to the Montserrat Company,
Montserrat, W.I.

THE leaves and young shoots of the lime tree yield a fragrant yellow oil
by distillation with water or in a current of steam. About 230 c.c.
were prepared for the purposes of the present investigation.

The oil is mederately soluble in rectified spirit of wine, and has a
sp. gr. = 0°8777 at 33° (water at 4° = 1). When agitated with soda
solution, no appreciable diminution of volume was observed.

The oil was submitted to fractional distillation, and the several
fractions were agitated with a strong solution of sodium bisulphite.
It began to distil at about 170°, and the portion which passed over
at this temperature did not combine with the bisulphite. It consisted
of a hydrocarbon. On the other hand, the fraction obtained at 220—
230° formed a crystalline mass with the bisulphite. But as this
fraction had no action or but very slight action on silver nitrate, it
was inferred that it consisted of a ketone rather than an aldehyde.

About one-third of the original oil boils at a temperature above
280°, has a viscid consistence, and exhibits a green fluorescence. It
has, in fact, the appearance of colophene, and may possibly have been
produced, at least in part, by the action of heat in the course of
distillation.

From the lowest fractions, after digesting with sodium, a hydro-
carbon was separated, boiling at 176—177°, inactive to polarised light,
and having a refractive index at 30° of 1°4611 for red light. This
compound unites with hydrogen chloride, forming a hydrochloride
which melts at 49—50° in an atmosphere of hydrogen chloride. It
also gives Riban’s colour reaction with ferric chloride. Except with
regard to its action on polarised light, this hydrocarbon resembles the
citrenes.

From the action of sodium on the lower fractions of the original oil,
and the production of a deep purple coloration when hydrogen

PERKIN: THE FORMATION OF ACIDS FROM ALDEHYDES. 317

chloride is passed into them, the presence of terpinol is rendered very
probable. It was not found possible to isolate it.

In order to determine the constitution of the ketone already referred
to, it was oxidised by means of chromic mixture. Abundance of
acetic acid was produced, and pelargonic acid was recognised by
means of its copper salt. These facts, coupled with the boiling point
(220—230°) of the compound, and the peculiar smell of pelargonium
oil noticeable in some of the fractions, prove that the substance is
methyl nonyl ketone, CH;CO-C,H)s.

The watery alkaline solution resulting from the first washing of
the original oil was found to contain acetic acid, and the liquid sub-
mitted to distillation smelt strongly of amylic acetate. The quantity
of this compound was, however, too small to admit of its separation.

From these results, it would appear that the principal constituents
of the oil of lime leaves are, a citrene (b. p. 176° circa), terpinol
(CyHi;"OH), methyl nony] ketone, and a colophene.

XXXIII.—The Formation of Acids from Aldehydes by the Action of
Anhydrides and Salts, and the Formation of Ketones from the Com-

pounds resulting from the Union of Anhydrides and Salts.

By W. H. Perkin, Ph.D., F.R.S.

THe reactions which take place when aldehydes are heated with
anhydrides and salts of fatty or bibasic acids, an account of which I
had the honour of bringing under the notice of this Society some
years since, have caused much work to be done with the object of
obtaining a clearer insight into the way in which they take place.
Fittig and his pupils, especially, have paid much attention to this
subject, and this has resulted in the discovery of many very in-
teresting products, and furnished us with fresh information respecting
the reaction itself.

Since I brought this subject before the Society, much knowledge
has been obtained in reference to the formation of condensation pro-
ducts, and it has been shown that the first phase of the reaction is,
in very many cases at least, analogous to that of the formation of
aldol from aldehyde discovered by Wiirtz in 1873; the splitting off
of water being a second result. Fittig has shown that a change of a
similar kind takes place in the formation of the unsaturated acids
which are produced by this reaction, and that their formation is pre-
ceded by that of hydroxy-acids or their derivatives containing acid
radicals, or of lactonic acids.

318 PERKIN: THE FORMATION OF ACIDS FROM ALDEHYDES

In my experiments, I was led to conclude that, in this reaction,
condensation takes place between the aldehyde and anhydride, and
Tiemann and Kraaz (Ber., 15, 2061) who, have worked a good deal
on the subject, are of the same opinion. Fittig and others, however,
consider that this is not the case, but that the condensation takes
place between the aldehyde and the salt. As the evidence they give
in support of this view does not appear to me to be conclusive, I have
taken up the subject again. My experiments, however, are not yet
completed, but as they have led to some new results which I desire
to lay before the Society, I have ventured to refer to some of them,
and to briefly consider their bearing. A very important point in the
discussion of this matter is the double decomposition which may take
place on heating the anhydride of one acid with the salt of another
acid: results have been obtained in reference to this subject, but the
work is still incomplete. I may state, however, that when sodium
isobutyrate is boiled with acetic anhydride, the compound of sodium
acetate and acetic anhydride I described some years since (Chem. Soc.
J., 1868, 21, 185) is formed, and therefore there can be no doubt about
double decomposition taking place in this instance. This is corrobo-
rated by Fittig and Otto (Annalen, 227, 79), who show by their
experiments that this takes place even at 100°. Sodium succinate
when boiled with acetic anhydride was also found to form the com-
pound of sodium acetate and acetic anhydride, and succinic anhydride
was obtained from the mother-liquors.

In a former paper on the subject of this reaction, I referred to the
fact that benzaldehyde with acetic anhydride yielded benzylidenz
diacetate, and said : “ It is possible that the formation of this substance
precedes that of the acid,” meaning cinnamic acid (Chem. Soc. J.,
1877, i, 425), and I am still inclined to believe that this is the case,
and that in the formation of other acids also, analogous products are
first produced from the aldehydes.

There is the remarkable circumstance in connection with the
formation of cinnamic acid, that the compound which (judging from
Fittig’s results) is the first product of the reaction, te. acetyl-
phenyllactic acid, has the same composition as benzylidene diacetate,
only a slight intramolecular rearrangement of the latter substance
being necessary for its formation :—

O-COCH; O-COCH;

| |
C.Hs CH C.H;-CH:CH,-COOH

Acetylphenyllactic acid.

BY THE ACTION OF ANHYDRIDES AND SALTS, 319

This acid, as shown by the experiment of Fittig and Slocum (Annalen,
227, 60), splits up when heated, yielding cinnamic and acetic acids,

O-COCH;

C.H,-CH.CH H:COOH = C,H,CH : CH-COOH + HOCO-CH,,

the reaction taking place more easily in the presence of acetic
anhydride.

The succinate of benzylidene does not appear to have yet been
isolated, but there is no reason to doubt its existence. Its composi-
tion is the same as that of phenylparaconic acid, and its change into
this compound may be thus written :—

| CO-OH
CH; CH but = CHyCH--OH

on, bu,

6—to o—tbo

Succinate of benzylidene. Phenylparaconic acid.

This acid on being heated also gives up carbonic anhydride, and
phenylerotonic acid is formed.

Why an acetate or other analogous salt should induce this intra-
molecular change in the benzylidene derivative is difficult to under-
stand. It is possible that the two first form a compound, and the
reaction may in some way be connected with the fact that the above
salts containing feeble acids are capable of furnishing a base for the
new acid. Thus in the formation of cinnamic acid I have shown that
the product of the reaction contains a compound of sodium cinnamate
and cinnamic acid.

Caro’s method of formation of cinnamic acid, which is used techni-
cally, also bears upon this subject. It is well known that silver
acetate and benzylidene dichloride yield benzylidene diacetate; if,
however, sodium acetate is used, this acetate, although undoubtedly
formed at first, yet changes directly it comes in contact with the
sodium acetate, cinnamic acid being formed as the final result. I
hope to make further experiments on the salts of benzylidene in re-
ference to this subject.

These observations point to the condensation taking place between
the aldehyde and anhydride, as do several of my older experiments.
Thus I found that cinnamic acid could be formed whether the salt
used were an acetate, butyrate, or valerate, but on these I would not

320 PERKIN: THE FORMATION OF ACIDS FROM ALDEHYDES

lay much stress, because it may be I was practically working to a large
extent with an acetate, the acetic anhydride entering into double de-
composition, butyric or valeric anhydrides and the salt being formed.
And as the acetic anhydride was somewhat in excess, and the more
active body at the temperatures employed, cinnamic acid would be pro-
duced, the other anhydrides taking no part in the reaction. But I also
made an experiment which is the opposite of this, and has moreover
been confirmed by the work of Conrad (Annalen, 204, 188), and on this
I do lay stress. I found that isophenylcrotonic acid could be formed
from benzaldehyde and propionic anhydride apparently equally well
whether sodium acetate or propionate were employed. — But, as re-
marked by Tiemann and Kraaz (Ber., 15, 2061), anhydrides of
acids richer in carbon would be feebler than those containing less
carbon ; therefore, acetic anhydride would turn out a compound richer
in carbon from its sodium salt. It would be very improbable that
the converse of this should take place to any considerable extent, and
in fact, experiment goes to bear out this view.

I have, however, made some more experiments on this subject. It
will be remembered that I found that when isobutyric anhydride was
heated with benzaldehyde and an isobutyrate, instead of obtaining an
unsaturated acid, carbonic anhydride and butenylbenzene were formed.
Tt appeared to me, therefore, that this anhydride might be used in
these experiments, because the carbonic anhydride given off might
serve as a measure of the extent of the reaction. Fittig has shown
that the first part of this reaction consists in the formation of
phenylhydroxypivalic acid (phenylhydroxytrimethacetic acid) or its
isobutyl-derivative, which are not very stable compounds, but when
heated split up into carbonic anhydride and butenylbenzene. This,
of course, does not militate against the following results, because the
final products still show the amount of change which has taken place in
the above direction. A mixture of benzaldehyde, isobutyric anhydride,
and an acetate was therefore experimented with. Now if the salt be the
agent with which the condensation takes place, cinnamic acid should
be produced, and no carbonic anhydride; but if the anhydride, then
carbonic anhydride and butenylbenzene should result. On heating
this mixture, carbonic anhydride was evolved freely and in abundance,
and butenylbenzene was also formed in quantity. In one experi-
ment, the amount of carbonic anhydride obtained in relation to the
aldehyde used was 24°2 per cent., and in a second experiment 22°5.
The temperature of the oil-bath used was about 170°, and the
time of digestion five and a half hours in one case, and six hours in
the other. A small quantity of cinnamic acid was formed. A mixture
of isobutyric anhydride, sodium isobutyrate, and benzaldehyde was
then taken, to see what amount of carbonic anhydride this would

BY THE ACTION OF ANHYDRIDES AND SALTS. 321

evolve in comparison with the preceding under similar conditions.
The temperature of the oil-bath was 170°, and the heating was con-
tinued for six hours. The amount of carbonic anhydride formed
amounted to 28 per cent. An operation was also made with iso-
butyric anhydride, sodium butyrate (normal) and benzaldehyde, a
temperature of about 160° being used, and the digestion continued
for five hours. This gave 21°6 per cent. carbonic anhydride, and only
17 per cent. phenylangelic acid.

In the above experiments, the proportions of the sodium salts and
aldehyde used were molecular, the amount of anhydride being rather
in excess of this proportion.

Three experiments were then made the reverse of the foregoing,
that is to say, acetic anhydride and an isobutyrate were used with
the benzaldehyde, molecular proportions being taken. These were
not expected to give any very definite results, because double de-
composition would be sure to take place between the anhydride and
salt, so that it would be a queston as to the rate at which this woald
take place, compared with that of the combination of the anhydride
and aldehyde. The results, however, show that cinnamic acid is rro-
duced in considerable quantity, and althongh carbonic anhydride
and butenylbenzene are formed, their amount is comparatively small.

The following are the results :—
II. IIT.

43-6 45°0
11-0 74

The temperature at which the last experiment was made was about
160°, the others nearer 170°. Q

The largest yield of cinnamic acid I have obtained by heating
acetic anhydride, an acetate, and benzaldehyde, was 84 per cent.
on the benzaldehyde, so that these experiments have yielded over
50 per cent. of that obtained under the most favourable circum-
stances.

Fittig and Ott made experiments with a similar mixture, but used
a temperature of 100° C., which had been previously shown by Fittig
and Slocum (Annalen, 22'7, 59) was too low to allow of cinnamic acid
being formed under any circumstances. The formation of this acid,
therefore, being prevented, phenylpivalic compounds, which are pro-
duced at this temperature, were gradually formed from the isobutyric
anhydride which was slowly produced by double decomposition. This
hypothesis is supported by the fact that when Fittig and Ott employed
acetic anhydride, the formation of phenylpivalic derivatives took
place very slowly, and only a small quantity was produced by heating
at 100° for 20 hours, whilst when isobutyric anhydride was used the
formation took place very much more quickly (Annalen, 227, 7%).

VOL. XLIX. Z

322 PERKIN: THE FORMATION OF ACIDS FROM ALDEHYDES

Other experiments are in progress in reference to this subject, but
I now wish to refer to some results which have been obtained while
studying the nature of this reaction, though they are not intimately
connected with it.

While examining into the changes which take place when normal
butyric anhydride, a butyrate, and benzaldehyde are heated together,
a high temperature being used (the mixture being in fact boiled), it
was noticed that carbonic anhydride was very slowly but continuously
given off. At first it was thought that this might be due to the decom-
position of the phenylangelic acid produced, but not being satisfied on
this point, it was determined to make experiments with butyric anhy-
dride and a butyrate, omitting the benzaldehyde, and on doing so it
was found that carbonic anhydride was still evolved in a slow but
regular manner, and when the volatile products were distilled, the
distillate not only contained butyric anhydride, but also a neutral oil
possessed of great stability.

Larger operations were then made. At first the proportions of
anhydride to salt used were molecular, but afterwards an excess of
anhydride was employed. The mixture was heated in a fractioning
apparatus with a column about 2 feet long; this was connected with
a condenser and a receiver, and this again with a potash bulb filled
with sulphuric acid to observe the progress of the operations by the
bubbles of gas passing. In one case bromine and water was used, to
see if any absorbable gases were formed, but none were found to be
produced. The mixture was kept boiling by means of an oil or
fusible alloy bath, but only gently ; a thermometer with its bulb in the
mixture generally jndicated a temperature of between 185° and 190° ;
this was maintained for about 36 hours, a small quantity of product
being distilled over now and then. The temperature was then in-
creased, so that nearly all the volatile products passed over. No
appreciable increase in the formation of carbonic anhydride was
observed during the extra heating, in fact it nearly ceased towards the
end of the distillation.

The product when fractioned showed no definite boiling point, it
was therefore treated with aqueous potash to remove butyric
anhydride or acid, usually present in the anhydride. This, how-
ever, was not found sufficient, and as the oil did not appear to be
decomposed by alcoholic potash, it was mixed with this in excess,
and, after standing for some time, separated by the addition of water,
and dried over potassium carbonate. On being fractioned, it com-
menced to boil at about 115°, the last parts not coming over until about
170° or 180°. After repeating this operation a few times, a boiling
point was indicated between 140° and 150°, and eventually the principal
part of the oil was obtained, boiling between 144°3° and 146°3°.

st fe ah hh elCUe, ak ele [Ue

BY THE ACTION OF ANHYDRIDES AND SALTS. 323

This gave on analysis the following numbers :—

[. Boiling point, 144°3—146°3°.
01433 gram substance gave 0°3885 gram CO, and 0°1592
gram OH.
II. Boiling point, 144—145°.
01458 gram substance gave 0°3937 gram CO, and 0°1617

Calculated for
I II. C,H,,0.

73-92 73°51 73°68
12:34 12:31 12°28

This substance is in fact butyrone, or dipropy! ketone.
The density determinations are pretty close to those given by
Chancel. He gives 0°8195 at 20°. The above gave—
= 0°83048

—. = 0°82165

d = 0°81452
0

Having only a limited quantity at my disposal, I do not assume it
was quite pure, especially as these substances, having a very low
specific heat, are very difficult to purify by fractioning, therefore the
density must not be taken as absolutely correct for pure butyrone.
This ketone is a very stable compound, and dissolves in concentrated
sulphuric acid with heat, apparently combining with it, but on the
addition of water it separates unchanged.

Isobutyric anhydride and sodium isobutyrate react in a similar
manner, producing a neutral oil consisting of various ketones, but
their examination has not yet been completed.

Having obtained the above results, similar experiments were made
with propionic anhydride and a propionate. On boiling these pro-
ducts together, carbonic anhydride is given off, but very slowly indeed,
the boiling point of the mixture being scarcely sufficient for the
reaction. The mixture was therefore heated in sealed tubes at 180—
190°. The amount of carbonic anhydride evolved was then sufficient
to burst some of the tubes, in fact it was found best to open them
after heating for three or four hours, and then to reseal and re-heat
them. After the tubes were opened and the carbonic anhydride
allowed to escape, condensing tubes were fitted to them, and they
were heated in an air-bath up to about 200°, until no more fluid came
over. This distillate was heated with aqueous potash, dried, and dis-

324 PERKIN: THE FORMATION OF ACIDS FROM ALDEHYDES

tilled ; most of it came over between 99° and 109°, leaving some
higher product; this was chiefly propionic anhydride, which had
escaped the action of the aqueous potash. On refractioning the
product, it again commenced to boil at 99°, but soon rose to 102°,
when most of it came over between this and 103°. This substance
gave on analysis the following numbers :—

0°1472 gram substance gave 0°3770 gram CO, and 0°1564 gram OH,,

Found. Calculated for C;H,,0.
69°76
11°62

This product was therefore propione. This ketone dissolves in con-
centrated sulphuric acid, and separates unchanged on the addition of
water.

On preparing a quantity of propionic anhydride from sodium
propionate and propionic chloride, and fractioning the product, a small
quantity of oil of low boiling point was obtained. When purified, this
boiled at the same temperature as propione, and had all its properties.
This would be most probably produced by the action of the anhydride,
as it was formed on some of the propionate which had escaped the
action of the chloride, though it is just possible it might have been
formed by the action of heat on some of the unchanged propionate
during the process of distillation; the latter, however, is the less
probable of the two.

Acetic anhydride was next heated with sodium acetate. No car-
bonic anhydride was given off, and on heating strongly acetic anhy-
dride distilled over, leaving acetate of sodium behind. If, however.
the mixture is heated in a sealed tube at 190—200° for three or four
hours, carbonic anhydride is formed in quantity, and the product
becomes brownish in colour. On opening the tubes and distilling off
the volatile products, a colourless liquid was obtained, evidently
containing acetic anhydride. On fractioning, a good deal passed
over between 60° and 120°, and this on being refractioned gave a
product boiling chiefly between 56°5° and 59°5°, which was evidently
acetone, as it combined with sodium bisulphite, forming the ordinary
easily soluble crystalline compound. It was not thought necessary to
analyse it.

This reaction therefore appears to be general between the salts and
anhydrides of the fatty acids. How can it be explained? Of cour-e
it is easy to understand that an anhydride, on losing carbonic anhy-
dride, can yield a ketone, thus :—

Be>0 = R>C0 + CO,,

BY THE ACTION OF ANHYDRIDES AND SALTS. 225

but an anhydride at the temperatures which have been used certainly
does not undergo this change.

Again, a salt of a fatty acid when heated yields, as is well known,
a carbonate and a ketone, but this decomposition requires a much
higher temperature than was used in the foregoing experiments, in
fact it was found that as soon as the mixture of the anhydride and
salt had been heated sufficiently high to distil off all volatile products,
no more carbonic anhydride or ketone was formed, so that this reaction
does not explain the matter.

In a paper published in this Journal a good many years since (1868,
p- 185), I showed that sodium acetate combined with acetic anhy-
dride to form the compound NaC,H;0,,C,H,O;, already referred to,
and there is no doubt that similar compounds are formed with other
anhydrides and salts: in fact, when they are heated together the salts
swell up, and a good deal of the anhydride disappears, and it seems
to me that the production of the ketones is due to the formation and
subsequent decomposition of such products, which are doubtless true
chemical compounds.

It is well known that benzaldehyde combines with acetic anhydride
to form benzylidene diacetate, which has already been referred to :—

H CO-CH 0-CO-CH
C,H,> 29 + 0<¢O-CH, =C.Hs'CH<6.60.CH,;

Acetaldehyde also combines with acetic anhydride to form ethylidene
diacetate.

If we apply this reaction to sodium acetate, which is simply alde-
hyde with an atom of hydrogen replaced by NaO, we get the follow-
ing result :—

NaO.~ ~ CO-CH, _ NaO 0-CO:CH
cH? ©? + °<co.cH; = CH.>C<0-CO-CH;

That this is the constitution of the compound NaC,H,;0-C,H,0,,
there is no reason to doubt, any more than in the case of the com-
pound obtained from oil of bitter almonds and acetic anhydride.
Now if this be the case, it is not difficult to see how this compound
should yield a ketone and carbonic anhydride—

O -CO-CH; _ .
©<oloo-cm, = CO + COLCH) + MeOOOOM,

becanse it appears quite probable that the weakest part of the com-
pound would be between the oxygen-atoms associated with the one
carbon-atom and the acetyl-groups.

326 PERKIN: THE FORMATION OF ACIDS FROM ALDEHYDES

The production of butyrone may be represented in an analogous
manner :—

Na0 CO-C\H; _ NaO. ,0-CO-C,H;
H,>©9 + 9<¢o.c:H, = C,H,>°<0-C0-C,H,’

the latter ikea breaking up as follows :—
Nee

|_O H ;
oxo oo on™ = CO, + CO(C;H,), + NaO-CO-C,H,.

If this explanation be true, and the sodium salt actually takes a
definite part in the reaction, it is clear, if a salt were used con-
taining an acid differing from that from which the anhydride was
produced, that unsymmetrical ketones should be formed, because the
hydrocarbon radical of the acid should go to form part of the ketone.
This was put to the test, and sodium acetate was used in conjunction
with butyric anhydride, because if it entered into the reaction it
would produce a ketone containing methyl, which could be easily
separated from other kinds of ketones by acid sodium sulphite ;
sodium butyrate would also be formed.

The reaction in this case should be—

Nao CO-C,H, _ NaOw , 0-C0-C,H
H,>©° + [<¢0.c,H, = CH,>°<0-CO-C,H,"

and then—

7 = CO, + CH;CO-C;H, + NaO-CO-C,H;.

On heating this mixture of sodium acetate and butyric anhydride,
the reaction went on much in the same way as when sodium butyrate
was employed, carbonic anhydride being regularly given off, but
the combination of the salt and anhydride took place, apparently,
more rapidly, so that it was necessary to add an excess of the latter
to keep sufficient fluid present to allow the boiling to take place pro-
perly. The crystalline salt gradually changed in appearance, and
after about 24 hours the product became perfectly fluid. The vola-
tile products were then distilled off, leaving a mass of fused sodium
salt, which was found to consist chiefly of sodium butyrate. By
fractioning, the oily distillate was separated into three portions—

BY THE ACTION OF ANHYDRIDES AND SALTS, 327

B. p. = 98—140°
140—165
165° and above.

The two first were mixed with sodium bisulphite and allowed to
stand a few days. The first fraction gave a considerable quantity of
a crystalline compound, the second gave no crystals, but a small
quantity of it dissolved. The crystals and solution of bisulphite
were separated from the undissolved oil, rendered alkaline, and then
distilled until oil ceased to come over. This oil was collected, dried
over potassium carbonate, and fractioned; it commenced to boil at
about 87°, but soon reached 100°, and nearly all had come over when
the temperature had reached 108°; when refractioned, it commenced to
boil at about 80°, but soon reached 103° ; the fraction boiling between
this and 104°5° was selected for analysis. Although the yield of this
product was considerable for the size of the operation, yet I did not
have enough to fraction the product very thoroughly, and it is probable,
judging from the boiling point of the product, that a little acetone
was formed in the operation, and that a minute quantity of this was
still contained in the product, as the numbers are a little low.

0°1467 gram substance gave 0°3718 CO, and 0°1531 OH).

Calculated for
Found. C,H, 90.

69°12 69°76
Hydrogen..... 11°59 11°62

Not having sufficient to refraction, it was combined with sodium
bisulphite, with which it soon formed a crystalline compound. On
analysis, this product gave the following result :—

02958 gram substance gave 0°1066 Na,SO, = 11°68 per cent. Na.

The formula C;H,O-NaHSO;,30H,, which is the formula given for
the compound of methyl propyl ketone, requires 11°56 per cent. Na,
therefore there can be no doubt that this product was methyl propyl
ketone ; it could not be its isomeride diethyl ketone, as this only com-
bines with bisulphite with great difficulty.

The oil unacted upon by the acid sulphite consisted chiefly of
butyrone ; that this substance should be formed is a necessary conse-
quence of the reaction, because as the sodium acetate becomes con-
verted into butyrate, this reacts with the butyric anhydride, and
yields butyrone.

These results, therefore, appear to establish the view of the reaction
which is given above.

There is one matter worthy of notice in connection with the forma-
tion of ketones by this new reaction, that is the formation of com-

328 PERKIN: THE FORMATION OF ACIDS FROM ALDEHYDES.

pounds of lower and higher boiling point than the ketone which alone
should be produced, and this seems to become more marked as the
molecular weight of the compounds used increases. These products,
which consist of ketones, I have not as yet been able to separate in a
satisfactory condition for analysis, but hope to be able to do so.

The formation of higher and lower ketones also occurs when the
calcium or barium salts of the fatty acids are distilled; thus the
product obtained from calcium butyrate is said to distil between 100°
and 160°; in the case of barium propionate the variation in tempera-
is less, viz., 80—108°.

It was thought that by heating a mixture of an anhydride of a fatty
acid with a formate, an aldehyde might be formed; it was found,
however, that formates decompose very readily under these circum-
stances, giving off carbonic oxide.

Experiments are in progress with reference to the action of the
anhydrides of bibasic acids on salts, but at present the results have
been unpromising: thus, when succinic anhydride is heated with an
acetate at about 190°, a black product is formed, which dissolves in
part in boiling water with a brown caramel colour, leaving a brownish-
black insoluble product.

ANNUAL GENERAL MEETING,

March 30th, 1886.
Dr. Hugo Miiller, F.R.S., President, in the Chair.

On behalf of the Council I have the honour of submitting to you
a report on the present state of the Society, and a brief account of
its activity during the past session.

It affords me great pleasure to say that those indications of pros-
perity which were dwelt upon by my predecessors in the more recent
Anniversary Addresses have not diminished, and that the increase in
the numerical strength of the Society is still fully maintained. This
will be borne out by the following statistics, quite apart from the
financial details, which I leave to be dealt with by your Treasurer.

Number of Fellows, Anniversary, March, 1885..... » 1360

Since elected and paid admission fees........ errr lll

1471

Re Ee ee Eee a a ie era
Withdrawn

Present number of Fellows

Increase 71.
Honorary Foreign Members at last Anniversary
Deceased

Total number of Members 1459

Thus, whilst the number of our members has increased to 1459,
the loss we have sustained by death during the year is somewhat
smaller than in the previous year. Amongst those, however, whom we
have to regret are some of our oldest and most prominent members.
The names of the deceased are; the honorary foreign member,
Prof. Hermann von Fehling, Frederick Allen, Prof. Andrews, E. O.
Brown, G. W. H. Clements, Frederick Field, F.R.S., James Higgin,
Rev. S. Morris, James Parette, Thomas Spencer, Dr. E. G. Schweitzer,

330 ANNUAL GENERAL MEETING.

Surgeon-Major Sheppard, Henry Shepheard, Alfred Tribe, Dr. John
Smith, W. S. Ward, and Walter Weldon, F.R.S.

The obituary notices will be given hereafter.

Passing on to the consideration of what must always be looked
upon as the principal function of our Society, it is particularly grati-
fying to notice the very marked increase in the number of the com-
munications and papers during the past year.

The following are the titles of the papers in the order in which
they were received :—

Papers contributed to the Chemical Society between March 30th, 1885,
and March 30th, 1886.

April 2nd.

1. “On the Formation of Hyponitrites from Nitric Oxide:” by
E. Divers and T. Haga.

II. “On the Existence of Barium and Lead Nitrosulphates :”
by E. Divers and T. Haga.

III. “ Notes on Fractional Distillation in a Current of Steam :”
by M. E. Lazarus.

IV. “ Derivatives of Taurine.” Part I: by J. W. James.

V. “On the Preparation of Ethylene Chlorothiocyanate and
8-Chlorethanesulphonic Acid :” by J. W. James.

VI. “The Orthovanadates of Sodium and their Analogues :” by
Harry Baker.

April 16th.
VII. “A Crystalline Tricupric Sulphate:” by W. H. Shen-

stone.

VIII. ‘*A Modified Bunsen Burner :” by W. H. Shenstone.

1X. “ The Chlorination of Phloroglucol:” by C. S. 8. Webster.

X. “Note on the History of Thionyl Chloride:” by C. Schor-
lemmer, F.R.S.

XI. “On the Reactions of Selenious Acid with Hydrogen Sul-
phide, and of Sulphurous Acid with Hydrogen Selenide :”
by E. Divers and T. Shimidzu.

XIT. **On a New and Simple Method of Quantitative Separation
of Tellurium and Selenium:” by E. Divers and M.
Shimosé.

May 7th.

XIIT. “On some Points in the Composition of Soils; with
Results illustrating the Sources of Fertility of Manitoba

ANNUAL GENERAL MEETING. 331

Prairie Soils :” by Sir J. B. Lawes, Bart, LL.D., F.R.S.,
F.C.S., and J. H. Gilbert, LL.D., Ph.D., F.R.S., V.P.C.S.

XIV. “ Researches on the Relation between the Molecular
Structure of Carbon Compounds and their Absorption
Spectra :” by W. N. Hartley, F.R.S.

XV. “ Researches on the Action of the Copper-zinc Couple on
Organic Bodies. Part X. Benzyl Bromide:” by J. H.
Gladstone, F.R.S., and Alfred Tribe.

XVI. “On the Selective Alteration of the Constituents of Cast
Iron :” by Thomas Turner, Assoc. R.S.M.

XVII. On the Existence of Nitrous Anhydride in the Gaseous
State :” by G. Lunge.

XVIII. ‘On the Reaction between Nitric Oxide and Oxygen
under Varying Conditions :” by G. Lunge.

May 2lst.
XIX. “A Colorimetric Method for Determining Small Quan-
tities of Iron :” by Andrew Thompson, M.A., B.Sc.
XX. “On some Sulphur Compounds of Calcium:” by V. H.
Veley.
XXI. “ Spectroscopic Observations on Dissolved Cobaltous

Chloride :” by W. J. Russell, F.R.S.

XXII. “On the Sulphides of Titanium :” by T. E. Thorpe,
F.R.S.

XXIII. “‘Note on the Formation of Titanous Chloride:” by
T. E. Thorpe, F.R.S.

June 4th.

XXIV. “On the Constitution of the Haloid Derivatives of
Naphthalene :” by R. Meldola.

June 18th.

XXV. “The Decomposition and Genesis of Hydrocarbons at
High Temperatures. I. The Products of the Manufac-
ture of Gas from Petroleum:” by Henry E. Armstrong
and A. K. Miller.

XXVI. “On the Non-crystalline Products of the Action of Dias-
tase upon Starch:” by Horace T. Brown and G. H.
Morris, Ph.D.

XXVII. “Decomposition of Carbonic Acid Gas by the Electric
Spark :” by H. B. Dixon, M.A., and H. F. Lowe, B.A.

XXVIII. “On the Influence of Silicon on the Properties of
Cast Iron: ” by Thomas Turner.

ANNUAL GENERAL MEETING.

XXIX. “Eleven Months’ Experience with Toughened Glass
Beakers:”’ by R. J. Friswell.

XXX. “ Bromo-derivatives of Diphenyl, Tolylpheny! and Ditoly|:
by T. Carnelley and Andrew Thomson.

XXXI. ** Note on the Influence of Strain on Chemical Action :”
by T. Carnelley and James Schlerschmann.

XXXII. “ On the Non-existence of Gaseous Nitrous Anhydride: ”
by William Ramsay, Ph.D.

XXXITI. “ On the Causes of the Decrepitation in Samples of so-
called Explosive Pyrites:” by B. Blount.

XXXIV. “ On the Specific Action of a Mixture of Sulphuric
and Nitric Acids upon Zine in the Production of Hy-
droxyamine :” by E. Divers and T. Shimidzu.

XXXV. “On the Action of Pyrosulphuric Acid on certain
Metals :” by E. Divers and T. Shimidzu.

XXXVI. “ On the Constitution and Reactions of Liquid Nitric
Peroxide :” by E. Divers and T. Shimidzu.

XXXVII. “ On the Behaviour of Stannous Chloride towards
Nitric Oxide and towards Nitric Acid:” by E. Divers
and T. Haga.

XXXVIII. “ Preliminary Note on the Reactions between Mercu-
rous Nitrate and Nitric Oxide, and between Mercurous
Nitrate and Nitrite: by E. Divers and T. Haga.

XXXIX. “ On some Derivatives of Anthraquinone:” by A. G.
Perkin and W. H. Perkin, Jun., Ph.D.

Papers Contributed to the Society during the Months of July, August,
September and October.

XL. “ On the Unit adopted for the Atomic Weights:” by
Lothar Meyer and Karl Seubert.

XLI. “* The Atomic Weight of Silver, and Prout’s Hypothesis :”
by Lothar Meyer and Karl Seubert.
XLII. ** A Method for Obtaining Constant Temperatures: ’’ by
W. Ramsay, Ph.D., and Dr. Sydney Young, D.Sc.
XLIII. “ Researches on Secondary and Tertiary Azo-com-
pounds.” No. III: by R. Meldola.

XLIV. “ Note on the Spontaneous Polymerisation of Volatile
Hydrocarbons at the Ordinary Atmospheric Tempera-
ture :” by Sir Henry E. Roscoe, F.R.S.

XLV. “ On the Action of Gypsum in Promoting Nitrification :”
by R. Warington.

XLVI. “Contributions towards the History of Formyl and Thio-
formyl Compounds derived from Aniline and Homologous
Bases :” by Alfred Senier, M.D.

ANNUAL GENERAL MEETING. 333

XLVII. “ Action of Phenyl Cyanate on Polyhydric and certain
Mononhydric Alcohols and Phenols :” by H. Lloyd Snape,
B.Sc.

XLVIIL. “ Chemical Examination of the Constituents of Camphor
Oil:” by Hikorokuro Yoshida.

XLIX. “On the Synthetical Formation of Closed Carbon
Chains :” by W. H. Perkin, Jun., Ph.D.

L. ‘‘ Action of Sodic Alcoholates on Ethereal Fumarates and
Maleates:” by T. Purdie, Ph.D., B.Sc.

LI. “ Contributions to the Chemistry of the Cerite Metals.” III:
by B. Brauner, Ph.D.

LII. ‘“* A New Method of Preparing Aromatic Hydrocarbons :” by
Richard Anschiitz.

LIII. “‘ On the Decomposition of Aromatic Ethereal Salts of
Fumaric Acid:’’ by Richard Anschiitz and Quirin Wirtz.

LIV. “ On an apparently New Hydrocarbon from Distilled
Japanese Petroleum:” by E. Divers and T. Nakamura.

November 5th.

LV. “ The Influence of Silicon on the Properties of Cast Iron.”
Part Il: by Thomas Turner, Assoc. R.S.M.

LVI. “ Modifications of Double Sulphates:” by Spencer Umfre-
ville Pickering, M.A.

LVII. ‘* The Relation of Diazobenzene-anilide to Amidoazoben-
zene:” by R. J. Friswell and A. G. Green.

LVIII. “ An Examination of the Phenol Constituents of Blast
Furnace Tar obtained by the Alexander and McCosh
Process at the Gartsherrie Iron Works:” by Watson
Smith, J. F. H. Coutts and H. E. Brothers. Part I.

LIX. “ The Decomposition of Potassium Chlorate by Heat :” by
Frank L. Teed, D.Sc., F.C.S.

LX. “ Note on the Refractive Power of Metacinnamene (Meta-
styrole) :” by H. G. Madan, M.A., F.C.S.

November 19th.

LXI. “ Aluminium Alcohols. Part III. Aluminium Ortho-
cresylate and its Products of Decomposition by Heat :”
by J. H. Gladstone, F.R.S., and Alfred Tribe.

LXII. “ Notes on the Constitution of Hydrated and Double
Salts :” by Spencer U. Pickering.

LXIII. ‘“* Some New Vanadium Compounds :” by J. T. Brierley.

LXIV. “ On the Action of Phosphorous Pentachloride upon
Ethylic Diethylacetoacetate : by J. W. James.

334

ANNUAL GENERAL MEETING.

LXV. “ On the Vapour-pressures of Mercury :” by W. Ramsay,
Ph.D., and Sydney Young, D.Sc.

December 3rd.

LXVI. “On the Presence of ‘ Raffinose’ in Barley:” by C.
O’Sullivan, F.R.S.

LXVII. “ On the Sugars of some of the Cereals and of Germi-
nated Grain:” by C. O’Sullivan,’’ F.R.S.

LXVIII. “ Atomic Valency:” by S. U. Pickering.

December 17th.

LXIX. “ The Action of Steam on Carbonic Oxide:” by H. B.
Dixon.

LXX. “ On Multiple Sulphates:” by Miss E. Aston and S. U.
Pickering.

LXXI. “ On the Use of Ferrous Sulphate in Agriculture :” by
A. B. Griffiths, Ph.D.

LXXII. “On Phenyltribromomethane, C,H,;CBr; :” by Walter

H. Ince.

January 21st, 1886.

LXXIIT. “The Influence of Silicon on the Properties of Cast
Iron.” Part IIL: by Thomas Turner.

LXXIV. “ The Chemical Action of Pure Cultivations of Bacterium
aceti:” by Adrian J. Brown.

LXXV. “A Method of Separating and Estimating Zirconium :”
by G. H. Bailey, D.Sc., Ph.D.

LXXVI. “ Notes on an Analysis of Koppite:” by G. H. Bailey,
D.Se., Ph.D.

LXXVII. “ The Monobromophthalic Acids :” by G. Stallard.

LXXVIII. “ Benzoylacetic Acid and some of its Derivatives.”
Part IV: by W. H. Perkin, Jun., Ph.D., and A. Calman,
Ph.D.

LXXIX. “Mercury Sulphites and the Constitution of Sul-
phites:” by E. Divers, M.D., F.R.S., and T. Shimidzu,
M.E.

LXXX. “ Potassium Chlorate :” by Frank L. Teed, D.Sc.

LXXXI. “ Note on the Chemical Formula for Wool Keratine :”
by R. Lloyd Whiteley.

LXXXII. “Some Derivatives of Thiourea:” by George McGowan,
Ph.D., F.R.S.E.

ANNUAL GENERAL MEETING. 335

February 4th.

LXXXIII. “The Chemical Formula for Wool Keratine:’’ by
Edmund J. Mills, D.Sc., F.R.S.

LXXXIV. “ Methods of Bacteriological Research from a Biolo-
gist’s Point of View:” by E. Klein, M.D., F.R.S.

February 18th.

LXXXV. “ The Constitution of Undecylenic Acid as indicated by
its Magnetic Rotation ; and onthe Magnetic Rotation, &c.,
of Mono- and Di-allylacetic Acids and Ethylic Diallyl-
malonate :” by W. H. Perkin, Ph.D., F.R.S.

LXXXVI. “ Reactions supposed to yield Nitroxyl or Nitryl
Chloride :” by W. Collingwood Williams, B.Sc.

LXXXVII. “ The Condition of Silicon in Cast Iron:” by A. E.
Jordan and Thomas Turner.

LXXXVIII. “Certain Aromatic Cyanates and Carbamates:” by
H. Lloyd Snape, B.Sc.

LXXXIX. “The Oil obtained from Lime-leaves:” by Francis
Watts.

March 4th.

XC. “ A New Element: Germanium:” by Clemens Winkler.

XCI. ‘*The Influence of Temperature on the Heat of Chemical
Combination :” by S. U. Pickering, M.A.

XCII. ‘The Salts of Tetrethylphosphonium and their Decom-
position by Heat:” by E. A. Letts and Norman Collie
Ph.D.

XCIII. “ The Formation of Acids from Aldehydes by the Action
of Anhydrides and Salts, and the Formation of Ketones
from the Compounds resulting from the Union of Anhy-
drides and Salts :” by W. H. Perkin, Ph.D., F.R.S.

XCIV. “A New Method of Preparing Tin Tetrethide:” by
E. A. Letts and Norman Collie, Ph.D.

XCV. “Contributions to the History of Cyanuric Chloride :” by
Alfred Senier, M.D.

XCVI. “The Action of Naphthylamine on Cyanuric Chloride :”
by Harold H. Fries.

XCVII. “Sulphine Salts containing the Ethylene Radicle.
Part I. Diethylenesulphide-methyl-sulphine Salts:” by
Orme Masson, M.A., D.Se.

XCVIII. “Sulphine Salts containing the Ethylene Radicle.
Part II. On Dehn’s Reactions between Ethylene Bromide
and Alkyl Sulphide :” by Orme Masson, M.A., D.Sc.

ANNUAL GENERAL MEETING.

XCIX. “ The Identity of certain Mixed Ethereal Oxalates :” hy
L. Gordon Paul.

March 18th.

C. “The Combustion of Cyanogen:” by Harold B. Dixon,
M.A.

CI. “ Note on the Constitution of the Naphthalene-derivatives :”
by R. Meldola.

CII. “ Action of Ammonia on Chromyl Dichloride:” by S.
Rideal, B.Sc.

CIII. “ Note on the Estimation of Resin in Soaps:” by C. R.
Alder Wright, D.Sc., F.R.S.

CIV. “The Properties of the Nitrobenzalmalonic Acids :” by
Charles M. Stuart, M.A.

It will thus be seen that the number of papers communicated has
reached 104, which is not only a very great advance on the previous
year, when the number was only 67, but with the exception of the
session of 1880-81, when it was 113, it is the largest number
hitherto attained. These contributions embrace a wide range of
subjects, and most of the branches of our science are represented
amongst them. A glance at the list calls to mind the great interest
and appreciation with which many of them were received and dis-
cussed at the meetings. This increase in the number and importance
of the communications has rendered it necessary, in order properly to
get through the business at these meetings, to adhere more strictly
to the practice of giving abstracts only of the papers presented.

But not only is this revival of chemical research evidenced by the
number of papers, but also by the greatly augmented size of the
volume in which they are recorded; this is nearly one-third larger
than that of the previous year, and almost twice as large as that of
1883-84.

The fature will show whether this satisfactory result is merely
accidental, or whether it really indicates the wished-for awakening
of activity in the direction of chemical research.

I am inclined to think that it is due in no inconsiderable degree to
the fact that the various new laboratories which have been established
in the country during the last few years are now gradually entering
upon their full activity. With the extending opportunities which
will be afforded to the younger generation of chemists by these
laboratories, we may look forward with confidence to a continuous
and steadily increasing development in the pursuance of scientific
chemical work in this country.

ANNUAL GENERAL MEETING. 337

In the course of the year we have had one lecture—delivered by Prof.
Klein—on “Methods of Bacteriological Research from a Biologist’s
Point of View,” which proved both highly interesting and instruc-
tive. The chemical study of the life processes of these low organisms
has already afforded most valuable results, and, in conjunction with
furt. »r and closer examination of the phenomena connected with the
action of the ferments generally, will open out a vast field for chemical
investigation of both scientific and practical importance.

As the identification of the particular organisms which are concerned
in these complex processes becomes more and more of importance, a
thorough acquaintance with the methods of pure cultivation of these
organisms must be considered as of special value, and Dr. Klein has
rendered a much appreciated service to the Chemical Society by giving
us the benefit of his great experience, gained in his own work in this
direction.

The experiment of publishing in a separate form the “ Abstracts of
the Proceedings of the Society’ has now been continued for more
than a twelvemonth, and I think you will agree with me that the
result has realised the object which prompted your Council in deciding
upon its initiation. The utility of this publication is particularly
appreciated by our country members, who are thus kept in touch
with the doings of the Society. The Honorary Secretaries still con-
tinue to charge themselves with the editorship, and our best thanks
are due to them for the zeal and promptitude with which they dis-
charge this self-imposed duty.

The Publication Committee is fully alive to the desirability of
effecting the publication of the Journal with the least possible delay,
and in this respect some notable progress has been made, thanks to
the efforts of our Editors.

I have also to direct your attention to that part of our publication
which contains the Abstracts of Papers published in other Journals.
The enormous amount of matter which has to be abstracted and incor-
porated into this volume had at the end of last year got somewhat in
arrear, but thanks to the vigilance of the editors this difficulty has
been overcome, and the Abstracts are now up to date.

In connection with this, I have to acquaint you with a recent
resolution of your Council as to the Abstracts of Papers on Technical
Chemistry. Whilst it has been felt all along that these Abstracts, on
account of their brevity, only very imperfectly fulfilled their object, a
more liberal treatment and greater completeness were considered to be
incompatible with the means at our disposal. After mature consider-
ation, the Council decided upon the discontinuance of these Abstracts
in the manner in which they have hitherto appeared, and for the
future only such papers on Technical Chemistry are to be abstracted

VOL. XLIX. 24

338 ANNUAL GENERAL MEETING.

and published as involve a new process, or possess otherwise a special
scientific interest. Previous to the foundation of the Society of Chemical
Industry, it was a matter of importance that Abstracts of Technical
Chemical Papers should, however short, find a place in our Journal,
but since the above-named Society has made it a special object to
publish in their excellent Journal comprehensive and full abstracts of
all technical chemical papers, as well as of patents, it was thought
that for the future we could not do better than leave this matter in
such competent hands.

Ever since the Chemical Society took possession of its present
home the extension and improvement of the library have been one of
its principal aims, and the Committee appointed for this object have
done their utmost to apply the means placed at their disposal to the
best advantage. Thus from year to year considerable and important
additions have been made, which are rapidly filling up the empty
spaces formerly so conspicuous in the bookshelves.

You will have noticed that a very considerable addition of space for
books has lately been gained by the new bookcases fitted up in the tea
room, and I may remark that these have been put up at the expense
of the Society, whilst the others in the library were taken over with
the premises as fixtures, and belong therefore to the Government.

In the last Anniversary Address, it was mentioned that the cata-
logue of the library which was then in progress was nearly com-
pleted, and I have now the pleasure of announcing to you that this
laborious task has just been finished, thanks to the assiduous energy of
Messrs. McLeod and R. Warington, and of your librarian, Dr. Thorne.
In the course of this work, it became evident that it was absolutely
necessary to check the presence of each volume enumerated in the
previous catalogue, and this, quite apart from the actual preparation
of the catalogue itself, involved a vast amount of labour, which has
been the main cause of the long delay.

The catalogue is now ready, and is being distributed to the members.
Your Council trust that its improved arrangement will meet with the
approval of the Fellows who are in the habit of using the library.

This may be deemed a fitting occasion for bringing under your
notice a statement drawn up by Dr. Thorne, which shows the addi-
tions to the Library made under the various headings, as far as the
present rearrangement and classification of the books permit a com-
parison with the account given three years ago.*

* In this statement the apparent decrease in “Technical Chemistry” is due to
the removal of pharmaceutical and medical works to a separate section, and of some
few other works to other sections. Works on meteorology are now included under

“Physics ;” and dictionaries are placed under the various sections to which they belong.
The decrease in the number of vo/umes of systematic works from 2600 to 2420

ANNUAL GENERAL MEETING.

1883. 1886.
Physics ... 255 288
General Chemistry 256
Inorganic Chemistry
Organic Chemistry
Physiological Chemistry ..
Vegetable Physiology and
Agriculture
Analytical Chemistry ....
Technical Chemistry
Mineralogy and Geology. .
Meteorology

ee ee
Pharmaceutical and Medi-

1803 (increase 28 4)

Volumes of systematic
WH nec cecanceews 2600
Volumes of Journals, &c.. 3600
Volumes of Duplicate Jour-
nals, &c., for circulation 600 839

8000 (increase 1200)
Pamphlets 76 1380 (increase 620)

Satisfactory as the steady growth of the library undoubtedly is,
there is, however, a difficulty looming in the distance; for as we may
hope that this development will continue, the space at our disposal
will become less and less, and very shortly the question of providing
fresh accommodation will arise.

As an event which has taken place during the past year, and which
is of special interest to this Society, I must not omit to mention the
Incorporation by Royal Charter of the Institute of Chemistry.

It will be within the recollection of some of those present that
about ten years ago a very strong desire was manifested by some of
our members that this Society should assume an authoritative posi-
tion for regulating and advancing the professional status of chemists,

is probably due to many of the old and short sets of periodicals having in 1883 been
counted amongst the volumes of systematic works. This would to some extent
account for the abnormally large increase (1750) in “ Volumes of Journals, &c.,”

although this is due in part to the completion of imperfect sets.
2a2

310 ANNUAL GENERAL MEETING.

with the view of ultimately obtaining the official recognition of
chemistry as a profession. This movement caused at the time a good
deal of perturbation, and seriously engaged the attention of those
who took a hearty interest in the welfare of this Society; but after
earnest consideration of the various means proposed for carrying out
the object desired, it was found that the Chemical Society was de-
barred by the articles of its Charter from giving effect to it.

Subsequently, an association was established with the special object
of bringing about a definite organisation of professional chemists,
and after having during the last eight years gradually developed its
aims and matured its plans, it has acquired an official acknowledgment
of its public utility by receiving a Royal Charter of Incorporation.

The Institute of Chemistry has thus entered upon a new career
which bids fair to confer a great benefit on all who are engaged in the
pursuit of chemistry in this country.

Regarding then from our Society’s point of view, the Institute of
Chemistry, in its present constitution, with its public duties and its
public responsibilities, I thmk we must rejoice that the originators
were forced by circumstances at an early stage of its initiation to
attempt the realisation oftheir laudable efforts on an independent
basis. It is quite inconceivable that the Chemical Society could
have taken charge of these new functions in their entirety without
seriously hampering the pursuits of its original and special object.
But whilst this issue in itself is a matter of congratulation, there is
still another aspect not less noteworthy, for we may anticipate that
the incorporation and official establishment of the Institute of
Chemistry will exert a direct and powerful influence on the develop-
ment of chemical education, which in its turn must assuredly promote
the progress of pure chemical science.

With regard to the operation of the Research Fund, I may state
that during the past year nine applications have been granted by
the Research Fund Committee, and the sums thus expended amount
altogether to £155, leaving even on the present somewhat reduced
income of £268 (of which sum £180 to £190 is derived from the
funded property) a very considerable balance, which in this instance
is perhaps not altogether a subject for congratulation.

As it might be expected from the foregoing statements that the
increased activity of the Society would have caused a proportionate
rise in the ordinary working expenses consequent upon the cost
of our enlarged pnblications, it is of importance to point out that we
have kept well within the means afforded by our income. For
reasons which will become apparent on a perusal of the financial
statement, last year’s expenses were actually somewhat less than in
the previous year, and consequently the increment in the income

ANNUAL GENERAL MEETING. 341

leaves us with a considerably larger balance, even if we make the
deduction for the probable expense of printing the catalogue which
has not yet been charged.

During the session ending March, 1884, the total expenses of the
Society were £2706, and the income £3556, leaving a balance of
£830 ; in the session ending March, 1885, the total.expenditure was
£3284, with an income of £3569, leaving a balance of £285. In the year
now ending the total working expenses of the Society amount to £3108,
whilst our income has been £3743; leaving thus a balance of £635.

In bringing this report to a close, I cannot refrain from expressing
the gratification I feel in having the privilege. of. testifying to the
manifest progress in the activity of our Society, and to the other-
wise prosperous condition of its affairs.

This is all the more encouraging when we look back upon those
years of its career when its activity seemed to. languish, and the
prospects of its development were not hopeful But some of my
predecessors who had occasion to refer to this state of things, and
who lamented over the apparent stagnation in the cultivation of
scientific chemistry in this country, when compared with what was
being done elsewhere, have perhaps not sufficiently recognised some
points which I think deserve to be more fully considered.

Thus it appears to me that sufficient cognizance was not taken of
the adverse influence of the eminently active life which surrounds us
here, and which lays everyone more or less under contribution for co-
operation in the attainment of that pre-eminence which this country
holds in commerce as well as in manufactures. And those who are
engaged in the pursuit of chemistry cannot help being drawn into
this all-absorbing vortex.

I will only refer to the vast amount of valuable work which is con-
tinually emanating from the large and important associations, the
Iron and Steel Institute, the Society of Chemical Industry, and also
in a minor degree, the Society of Public Analysts, each of which, in its
special line, represents chemical activity in this country.

Those who formerly entered on the study of chemistry, mostly did
so with a view to its ultimate practical application, and strove in
perhaps undue haste to attain this object. Under such conditions
individuals who had the qualification, the means, and the leisure to
devote themselves to scientific chemistry were naturally but few.

Now if we enquire into the causes which favoured the development
of scientific chemistry in so marked a degree in other countries, we
find that it took its rise in the seats of the higher education, in the
universities, where its study was carried on on a level with that of
other sciences, whilst its immediate utilisation for practical purposes
remained for a considerable period in abeyance, and, therefore, the

ANNUAL GENERAL MEETING.

material advantages accruing from it to the public at large were but
small. In the course of time, however, its general diffusion produced
its effect on the development of the arts generally, and gave more
especially a powerful impetus to the progress of technical chemistry
in its various branches.

This latter fact being now fully recognized in this country, great
efforts are being made to stimulate the cultivation of scientific
chemistry, and the establishment of late years of numerous labora-
tories throughout the country, to which I have already referred,
cannot fail, in due time, to contribute their share towards the general
advancement of our science, a result which this Society in particular
must hail with acclamation.

Mr. Frepertck ALLEN was born-at Dartford, Kent, in 1809. He
was educated privately, and apprenticed to an eminent firm of manu-
facturing chemists in London.

In 1833 he commenced business operations, devoting his attention
specially to the production of oxalic acid and the oxalates. Subse-
quently he took much interest in the manufacture of artificial manures,

and ata later period, in the early days of the coal-tar industry, was
among the first to produce nitrobenzole and aniline on a manufacturing
scale.

He was a thorough practical chemist, very energetic, and always
desirous of furthering the interests of chemical industry.

Tuomas ANDREWS was a native of Belfast, and was the son of a
local merchant. It was in this town he received his early training, and
here, too, in the old Belfast College, of which he afterwards became
one of the professors, he obtained his more advanced education ;
he also spent some sessions in Edinburgh University and Trinity
College, Dublin. Having completed his collegiate course, which was
a most successful one, and obtained his degree of M.D., he settled
down in Belfast as a medical practitioner. After the lapse of a com-
paratively short period, he was appointed to the Chair of Chemistry
there, and it may be remarked, as showing the high reputation which
he had acquired at that early stage of his professional career, that,
at the time of his acceptance of the professorship, he had at his
choice two lectureships—one in Dublin and the other in Edinburgh.
Dr. Andrews, however, chose to remain in Belfast, continuing his
practice as a medical man, and attending, at the same time, to his
professorial duties. Thus he continued up to 1846, when, the
founding of the Queen’s University in Ireland having been decided

ANNUAL GENERAL MEETING. 343

upon, he was selected as Professor of Chemistry and Vice-President in
Belfast Queen’s College—a dual office which he certainly adorned.
At the same time the Rev. Dr. Shuldham Henry, of Armagh, was
appointed President, and it was Jargely under the direction of these
two gentlemen that the College buildings were reared. President
Henry and Vice-President Andrews afterwards lived for many
years in their adjacent residences in the College .buildings, and both
resigned their respective offices at the same period. It was not until
1849 that the new College was opened, and then Dr. Andrews having
commenced in it his duties as Professor, began to devote himself
greatly to original research, especially as regards the specific heats
of different substances. He afterwards gave much time to researches
in reference to ozone, and in later years applied his great powers to
the condensation of gases. Two of his papers on the liquefaction of
gases were chosen as the Bakerian Lectures of the Royal Society.
The first of these was in 1876, and the second shortly afterwards, and
he had the high honour of reading both before the members of the
Society. Dr. Andrews never devoted much time to literary work,
but, in addition to several scientific papers, he published in 1868 a
pamphlet entitled “ Studium Generale ; or, a Chapter of Contempo-
rary History,” dealing with the question of Irish university educa-
tion. In 1876, however, when the British Association visited
Glasgow, Dr. Andrews was chosen to the distinguished position of its
president, and on that occasion he delivered a remarkably able and
exhaustive review of the advancement of science in general. In 1879,
in consequence of increasing infirmities, he resigned his official con-
nection with the Queen’s College, and was succeeded in the chair
of chemistry by the present occupant, Dr. Letts. He then went
to reside at Fortwilliam Park, and there he passed away after
a prolonged period of illness. Dr, Andrews was an honorary
LL.D. of the Universities of Dublin, Glasgow, and Edinburgh,
He was a Fellow of the Royal Society, Fellow of the Chemical
Society of London, and Honorary Fellow of the Royal Society of
Edinburgh. In the Queen’s College, he was a great favourite with
the students, who ever manifested a thorough appreciation of his
masterly lectures. Long after the students had completed their
training, he evinced the greatest interest in their welfare, and took
special delight in watching their advancement and prosperity. It
will be remembered that on Dr. Andrews’ retirement from the
College a life-size portrait of him was placed in the Common Hall,
and he was presented with a replica for his own residence. In addi-
tion to adorning the College walls with his likeness, further honour
was conferred upon the Doctor by the founding of the Andrews’
Scholarship. Dr. Andrews was married to a Scotch lady, Miss

844 ANNUAL GENERAL MEETING.

Walker, daughter of Major Walker, who survives him, and by whom
he leaves three daughters and two sons, one of whom is a major in
the army and the other a member of the Irish bar. It may be stated
that Lord O’Hagan and Dr. Andrews were most intimate friends.
They were schoolboys together, and the close friendship which was
then formed continued unbroken. Only two years last autumn, when
the late Lord O’Hagan himself was not at all well, he came to Belfast
with the express object of paying a visit to Dr. Andrews. That was
the last visit he paid to this town, to which he was so much attached.
The death of this eminent scientist will be received with deep regret,
not only in all parts of the United Kingdom, but thronghout the
scientific world.

Epwin Ormonp Brown was born in the year 1828, at Devonport,
where he served his time as an apprentice to a chemist and druggist.
From an early age, he exhibited much interest in scientific pursuits,
his leisure time being chiefly devoted to his favourite branches of
study, Chemistry and Electricity.

His associates (amongst whom were several medical students) fully
appreciated the talent he possessed of imparting to others in a clear

and concise manner the knowledge he had acquired, and with the
progress of time and increased experience this talent developed itself
more fully, as will be allowed by all those who have benefited by his
instruction.

In 1849 he entered as a student at the College of Chemistry under
Dr. Hofmann. Here he gained the character of being an earnest
investigator, eager to obtain a thorough knowledge of chemical
science, and displaying great originality of ideas when the various
theories propounded by eminent chemists came under discussion.
After leaving the College, he became Lecture-Assistant to the late
Dr. Stenhouse at Saint Bartholemew’s Hospital, and subsequently
assistant to Professor Way.

In 1856, he was appointed an assistant chemist of the War Depart-
ment. Some of his earlier work consisted in an investigation of com-
mercial coppers; during this he devised a more exact method of
determining copper than had been previously known, which was
published in the Society’s Journal, 1858, and still continues to be used
as one of the most exact and rapid methods of determining copper.

In 1862, the War Office determined that the question as to whether
gun-cotton could be manufactured in such a form as to be capable of
being used in the Service as an explosive agent should be thoroughly
investigated by Professor, now Sir Frederick, Abel, and during the
five years from 1863 to 1868, experiments were made on a somewhat
considerable scale at Waltham Abbey, until a safe and efficient

ANNUAL GENERAL MEETING. 345

process of manufacture had been established. These experiments
were conducted by Mr. Brown, and the fact that they were carried
out without a single accident of any importance speaks highly for the
care and talent he displayed. He continued to be employed in inves-
tigations connected with the applications of gun-cotton, as well as
other explosives, until his death.

Amongst the numerous valuable results obtained from his investi-
gations, one of the most important was the fact that gun-cotton when
converted into the pulped compressed condition as devised by Sir
Frederick Abel was susceptible of being detonated whether in a dry
or wet condition, by which the greatest amount of force it was
capable of developing could be very readily utilised.

In 1876, Mr. Brown undertook the teaching of the Instructors in
Gunnery of H.M. ship “ Exceilent,” as well as voluntary classes of
naval officers, in applied electricity, thus commencing a system of
instruction which afterwards developed into the establishment of a
Naval Torpedo School at Portsmouth. He also instructed classes of
Royal Artillery officers and cadets in applied electricity and in the
military applications of explosives.

Mr. Brown died very suddenly on the 5th December, 1885, after

a few days’ illness, no apprehension of his being in any danger

having been entertained by his medical attendant until the morning
of the day on which he expired.

Georce Wittiam Hotumay Ciements, F.C.S., was born at Carmar-
then, South Wales, in 1857. When ten years old he was sent to a
private grammar school, at St. Helen’s, Lancashire, conducted by
Mr. Thomas Bournes, under whose tuition he soon mastered the
simpler elements of chemical science. When scarcely 15, he
was apprenticed to Mr. J. A. Phillips, then Managing Director of
the Widnes Metal and Smelting Works, at Widnes, Lancashire.
Under the supervision of this celebrated chemist, he made rapid
progress in mastering the technicalities of the science, both in the
laboratory and by practical operations in the works. Here he con-
tinued for nearly five years, when a vacancy having taken place in the
Laboratory Department of the Rio Tinto Copper Mines, at Huelva, in
Spain, he was nominated to that post by Mr. Phillips, then Consulting
Engineer to the Company. Mr. Clements went out to the mines in
December, 1876, and for about two years (with entire satisfaction)
filled the position of assistant chemist there.

He returned to England at the end of 1878, and for some time was
engaged in analytical work in London, chiefly in the laboratory of
Messrs. Mattheson and Co., the former proprietors of the Rio Tinto
Mines. In the early part of 1880, he was offered the post of Analyst

346 ANNUAL GENERAL MEETING.

(and Assayer of Ores) to the Bilbao Iron Ore Company Limited, in
the north of Spain. For three years he held this responsible position,
until the company ceased mining operations.

For the two last years of his life, he practised in Bilbao solely on
his own account, as an analytical chemist and assayer of metallic ores.
Owing to his sound practical experience in chemical analysis, know-
ledge of mines and minerals in general, with a character full of
energy and enterprise, combined with high moral principles, he never
failed to secure the confidence and support of his patrons, and all
with whom he had intercourse. Had his life been spared there is
hardly any doubt but that he would have made his mark in inventive
science. In his spare moments, he projected many ingenious sug-
gestions in mechanical science, improvements in locomotive engines,
utilisation of tidal forces, &c., &c., which if spared, he hoped more
fully to develop.

He died at the early age of 27, at Cumino de Santioce, Portugalate,
Spain, on the 4th May, 1885, after a very brief illness, from conges-
tion of the lungs, and was interred in the English cemetery, Bilbao,
Spain. A monumental tablet was erected by express wish and at the
sole cost of the English residents there, as a mark of respect to his
memory. He has left a young widow and one posthumous child.

HERMANN VON Fete was born on the 9th of June, 1812, at Liibeck.
Originally intending to devote himself to pharmacy, he entered,
about the year 1835, the University of Heidelberg, where he worked
in Gmelin’s laboratory. After graduating at Heidelberg, he removed
to Giessen, where his earliest researches were carried out under
Liebig’s guidance. His first published work was a refutation of
EK. Davy’s statement concerning the isolation of fulminic acid.
During the Giessen period, he also investigated with success the
polymerides of aldehyde and the composition of the salts of sulpho-
benzoic acid.

Soon after completing his studies, Fehling was appointed, on
Liebig’s recommendation, to be Professor of Chemistry in the Poly-
technic at Stuttgart, a post which he held for more than forty years.
At first his time was fully occupied in organising the chemical de-
partment of the institution, but as soon as his duties permitted he
returned to the prosecution of his researches. His work on ammonia,
on the palladium-ammonia compounds, and his monograph on succinic
acid, date from this period. About this time, too, in 1844, Fehling
carried out one of his most important investigations—that which
led to the discovery of benzonitrile. This compound, which he
obtained by the elimination of the elements of water from ammonium
benzoate, was the prototype of the class of organic nitriles. It

ANNUAL GENERAL MEETING. 347

is interesting to note that Fehling was fully alive to the bearings
of this experiment; for not ouly does he, in stating his results, refer
to the analogous transformation of ammonium formate into bhydro-
cyanic acid discovered by Pelouze, but he calls attention to Dibe-
reiner’s almost forgotten experiment in which cyanogen is obtained
from ammonium oxalate.

Later on, Fehling deserted pure chemistry for. subjects of more
practical interest. His professorship at the Polytechnic School, and
the various other official appointments which he simultaneously held,
directed his energies into these new channels. Questions of tech-
nology and of public health mainly occupied his attention. A list of
the various subjects of this nature which he investigated could not be
compressed within the limits of this notice. He also devised several
new analytical methods, the best known of which is that for the
determination of sugars by means of the soiution which bears his
name.

In spite of the constant pressure of professional and other official
duties and of his experimental researches, he succeeded in accom-
plishing an amount of literary work which might have filled the
entire time of a less energetic man. The German edition of Payen’s
Chimie Industrielle is from his pen. He took an active part in
bringing out the later editions of Graham-Otto’s Teat-book of
Chemistry. He was a contributor to the Handwérlerbuch of Liebig,
Poggendorff, and Wéhler, and when the Newes Handwéorterbuch was
to be produced, Fehling undertook the task of editing it, a work
which he has not lived to complete. For many years he was a
member of the Committee appointed for the revision of the Pharma-
copwia Germanica.

He died on the Ist of July, 1885.

Freperick Fienp, F.R.S., F.C.S.E., Member of the Institute of
Santiago de Chile, &c., was born in Lambeth, August 2nd, 1826. He
was the second son by the second marriage of the late Charles Field,
partner in the old established firm of J. C. and J. Field, wax bleachers
and manufacturers of wax and sperm candles. He received his early
education partly at Denmark Hill Grammar School, and partly by
private tuition. He afterwards went to Mr. Long’s School at Stock-
well, in 1841, where he was a schoolfellow of Professor Odling.
Young Field left Stockwell at the end of a year, and having from
childhood evinced a great partiality for chemical science, he was
placed in the Laboratory of the Polytechnic Institution, in 1843, at
that time conducted by Dr. Ryan.

From the very popular nature of the Institution, Field's progress
was not satisfactory, he however did his best to acquire some practical

348 ANNUAL GENERAL MEETING.

knowledge of chemistry there, and was a diligent reader in his even-
ings at home of the works of Brande, Graham, Daniel, and others,
The late Mr. Mitchell, a consulting chemist of some repute, soon
afterwards induced Mr. Field to join him in his laboratory as an
assayer and analytical chemist, another ill advised step which he took,
as although he had spent much time in the assay of metals by the
“dry” process in the laboratory of the Polytechnic, he was far too
young and inexperienced to combat with the many difficulties with
which he was surrounded, and to make himself responsible for the
mercantile worth of valuable cargoes of mineral or other natural
products. His own sense of his want of greater knowledge led him,
as soon as the Royal College was established, to enlist himself as a
pupil under Dr. Hofmann, who was at that time ably supported by
the late Dr. Blythe. It was at this time that Mr. Field was elected a
Fellow of the Chemical Society (1846), and he soon after published
his first paper ‘*‘On the Products of the Decomposition of Cuminate
of Ammonia by Heat, and on Nitrobenzamide,” which was read before
the Society in 1847. There was a certain interest attached to this
paper, as “‘Cumonitrile,” one of the products of the decomposition

of cuminate of ammonia, was the second of that interesting class of
bodies yet discovered, Felling having previously described benzo-

nitrile.

In the autumn of 1847, Field was offered an appointment as chemist
to a copper smelting establishment about to be erected in the province
of Coquimbo, Chili, South America. Although very anxious to
pursue his researches under the able guidance of Dr. Hofmann,
domestic circumstances induced him to accept the offer, and he arrived
in Chili at the commencement of 1848. There was, of course, con-
siderable delay in erecting an establishment of the kind, but by dint
of great perseverance on the part of the superintendent, Mr. Alison,
the smelting of copper ores on a large scale commenced early in
1849, There is a paper by Field, in the Jowrnal of the Chemical
Society, 1850, “On the Examination of some Slags from Copper
Smelting Furnaces.” As the process of smelting used in Chili
differed somewhat from that in England, the results of the examina-
tion proved not unworthy of note. There is also another paper on
the ashes of the cactus, a plant which grows luxuriantly in these
climates, and which was of interest from the large amount of car-
bonate of soda it yielded, derived from the sodium oxalate which exists
abundantly in the plant. In 1851, Field published a description of a
natural alloy of silver and copper, which bore a great resemblance to
nearly pure silver, and proved sadly disappointing to the discoverers
of the mine. An interesting circumstance occurred soon after this.
A large quantity of ore arrived at the establishment, supposed to

ANNUAL GENERAL MEETING. 849

be the blue carbonate of copper (chessylite), not an: uncommon
mineral in the neighbourhood of Coquimbo. Great was the indigna-
tion of the importer when he was informed by Mr. Field that his
cargo did not contain a trace of the metal. It was, however, far too
interesting to pass unnoticed, and on a minute examination it was
found to be a pure lapis lazuli, the first found in South America, and
Field was able to prove conclusively, that the sulphur did not exist as
sulphuric acid, but as a sulphide, inasmuch as after digestion in
strong acetic acid and evolution of sulphuretted hydrogen no trace of
sulphur could be detected.

In 1852 it was proposed by the company in which Field was engaged,
to establish smelting works in Caldera, a new port north of Coquimbo,
on the confines of the desert of Atacama, in the neighbourhood of
a district peculiarly prolific in copper and silver ores, and he was
appointed manager of the new works. After an absence of five years
he revisited England, in November, 1852, and soon afterwards married
a sister of Sir Frederick Abel. He returned to Chili in the spring of
1853, and went at once to Caldera, which, although a new port, was
rapidly increasing in importance. A railway was already established
(the first on the western coast of South America) between the port
and the city of Copiapo, and every inducement was afforded to
foreigners, as well as the natives of the country, for the exploration
of that district, the trade of which soon became of such importance
that a British Consulate was deemed necessary, and Field had the good
fortune to be appointed Vice-Consul in 1853. The duties of this post
were not very arduous at certain seasons, but occasionally there was
sufficient work to occupy many hours each day, especially during the
period of the Russian War, when, owing to France not being repre-
sented, Field was desired by the late Admiral Harris, Chargé
d’Affaires in Santiago, to assume the duties of French Consul in
addition to those of England, and to afford the same protection to
Frenchmen as he would to those of his own nation. In 1856, Field
left the company in which he had been engaged for so long, and was
appointed as chemist and under manager to a smelting establishment
at Guayacan, which, though small in its beginning, is now perhaps
the largest copper works in the world.

In 1859, a revolution of very formidable proportions broke ovt in
Chili. The revolutionary party, headed by Don Pedro Gallo, marched
from Copiapo, and after some days’ journey arrived at Coquimbo,
where a battle was fought between them and the Government troops,
the latter being completely routed. As the establishment in which
Field was now manager belonged almost exclusively to Seiior
Urmeneta, a zealous partisan of the Government, his position was
very difficult and somewhat perilous; moreover, he had just been

350 ANNUAL GENERAL MEETING.

appointed Acting Consul by the Chargé d’Affaires, in the absence of
the Consul. Mr. Field, however, managed to send his wife and family
to England, and remained at his post until the revolution had sub-
sided, when he had the gratification of restoring the establishment,
almost intact, into the hands of its owner, after which he left Chili,
and arrived in England in September, 1859. He was appointed
Lecturer of Chemistry at St. Mary’s Hospital, in 1860, and Professor
of Chemistry at the London Institution in 1862. In the autumn of
that year, his old friend and fellow student, Mr. E. C. Nicholson,
then of the firm of Simpson, Maule, and Nicholson, offered him
a situation as chemist to their Aniline Colour Works, where he
remained until 1866, when he became a partner in the house of
J. C. and J. Field. The numerous duties incidental to so large
a business precluded him, daring the last few years of his life,
from devoting his energies to the pursuit of his favourite science.
His health, moreover, began to fail him very much in 1876, and he
had to pass two winters entirely away from England. For the last
two years of his life, he was a great and patient sufferer, and he died
on Good Friday, April 3rd, 1885, at the age of 58.

The following may be quoted as among the more important of
Mr. Field’s contributions to chemical science :—“ On the Products of
the Decomposition of Cuminate of Ammonia by Heat and on Nitro-
benzamide ;” ‘On the Composition of the Ashes of the Cactus,”
“Description of Lapus Lazuli, found in Large Quantities in the Cor-
dilleras of the Andes ;” “On the Composition of a Specimen of Ataca-
mite from the Province of Copaipo, Chili,” “On the Analysis of a
Surface Soil from the Desert of Atacama,” “On the Action of Heat
on Oxychloride of Copper,” “ On Algodonite, a New Mineral, contain-
ing Arsenic and Copper,” “On the Separation of Iodine, Bromine,
and Chlorine, with some Analyses of their Combinations with Silver
in Chili,” ‘“‘On. the Action of Hydrochloric Acid on Sulphate of
Mercury, in the Presence of certain other Substances,” “‘On Some
Minerals containing Arsenic and Sulphur from Chili,” “On Some
Native Combinations of Oxide of Mercury with Oxide of Antimony,”
“On the General Distribution of Bismuth in Copper Minerals,”
“Some Results of the Analysis of Commercial Coppers” (joint
paper with Abel), “‘On the Double Snlphides of Copper and Iron,”
** On the Existence of Silver in Sea-Water,” “On the Separation of
Iron from Manganese,” “On the Cuprous Oxide of Manganese from
Chili,” ‘“‘ On the Artificial Formation of Atacamite,” “On Various
Methods for the Estimation of Copper,” “ On the Double Arseniate of
Magnesia and Ammonia,” “ On Ludlamite, a New Mineral,” “On a
Brittle Variety of Silver from Bolivia,” &c.

— Oe 2. eee eee eee. ee ee Oe ee hd! tei ces ee

ANNUAL GENERAL MEETING. 351

James Hiaain was the fourth son of Thomas Housman Higgin, of
Greenfield, Lancaster, by his wife, Sarah, daughter of the Rev. James
Winfield, M.A., of Chester. He was of an old Lancashire stock, his
father’s family having been settled since the latter part of the
fifteenth century on a small property near the town of Colne, in
Lancashire. He was born in the year 1824, in Lancaster Castle, of
which his father was then governor. His education was commenced
at the Lancaster Grammar School, and subsequently he studied
under the Rev. James Wheeler. He left Lancaster in 1839, and
went to Horwich Vale Print Works (the property of his uncle,
Mr. Thomas Chippendale), where he remained until the business was
given up about 1846. He next went to Manchester, where he worked
in the laboratory of Dr. Angus Smith. While he was at Horwich
Vale, he commenced an elaborate investigation of the constituents of
madder. The work was finished during the time that he was with
Dr. Angus Smith, and in 1848 he published a paper in the Philoso-
phical Magazine (33, 282—294) with a view of showing that
Dr. Schunck’s assertion that madder owed the whole of its
tinctorial power to alizarin was incorrect. Dr. Schunck had de-
scribed madder as consisting “of three principles, xanthine, a
yellow; rubiacine, an orange; and alizarine, a red colouring matter,”
and said that the two former were valueless for colouring pur-
poses. Mr. Higgin showed that extract from madder undergoes a
remarkable spontaneous transformation when allowed to stand,
inasmuch as the alizarin increases in quantity owing to the decom-
position of the other constituents. He considered that this was a de-
velopment closely akin to fermentation, for he found that the change
proceeded most rapidly at temperatures from 120—130° F. He,
therefore, stated his belief in the theory advanced by Decaisne that
all the colouring matter of madder arises primarily from xanthin,
especially as fresh madder roots contain only a deep yellow fluid. It
has since been ascertained that madder contains a ferment to which
the name of erythrozym has been given, The publication of this
paper was the means of bringing Mr. Higgin into connection with
Dr. Schunck, who admitted the correctness of Mr. Higgin’s views, and
owned himself wrong. From this time, up to 1850, Mr. Higgin
remained in Dr. Schunck’s works at Rochdale, when his principal
study was the manufacture and use of tin salt. In 1850, he entered
into partnership with Mr. Nathan Lloyd, and carried on an extensive
business as a chemical manufacturer. On the retirement of Mr. Nathan
Lloyd, his brother Abraham took his place, and the business was
carried on under the name of James Higgin and Co., until 1874, when
it was changed to Higgin, Lloyd, and Co. Mr. Higgin’s practical
experience of the wants of dyers and calico-printers peculiarly fitted

852 ANNUAL GENERAL MEETING.

him for the work of manufacturing chemicals and dyes for them, and
his extensive knowledge and inventive genius combined to make the
business a success. Although Mr. Higgin did not publish many
papers, the amount of chemical work which he did was very large.
He was the author of the article on Calico Printing in Ure’s
Dictionary. In 1850 he published a paper in the Chemical Gazette
(8, 249—250), on a qualitative test for nitric acid. In 1881, together
with his son Alfred J. Higgin, he took out a patent for an electric
battery, in which tin is substituted for zinc. In this way, the refuse
from the battery can be used for making tin salt. Mr. Higgin died
on June 2lst, 1885, at his residence, The Hollies, Timperley, near
Manchester, and was interred at Kersal, near to his old friend
Dr. Angus Smith.

He was elected a fellow of the Chemical Society in 1852.

Mr. Atrrep TripeE was born in London in the year 1839, in
humble circumstances, and his first acquaintance with science seems
to have been obtained as a boy at the Royal College of Chemistry.
While waiting upon the students there, he acquired whatever know-
ledge he could, and repeated at home many of the experiments he
had seen them perform. Dr. Hofmann, pleased with his desire for
knowledge, gave him every encouragement and assistance in his
power. In 1855, at the age of 16, he became the assistant of
Dr. Medlock, then of Dr. Forbes Watson, and afterwards assisted
Prof. Williamson, of University College. He then went to
Dr. Bernays, who in 1862 induced him to spend a year at Heidelberg
under Prof. Bunsen, and kept his place at St. Thomas’s Hospital open
for him while he was away. On his return, he continued to act as
laboratory assistant and Demonstrator of Chemistry.

In 1865 he became the private assistant of Dr. Gladstone, and
remained head of the laboratory till his death. From 1877 he held
the Lectureship on Metallurgy to the Medical School of the National
Dental Hospital, and since 1874 he was Lecturer on Chemistry and
Director of Practical Chemistry in Dulwich College.

He became successively a Fellow of the Chemical Society, of the
Institute of Chemistry, and of the Society of Chemical Industry.

Mr. Tribe was pre-eminently a scientific investigator. He loved
patient and original research, and all his work was most carefully and
honestly done. He published a large number of papers principally
on the borderland between chemistry and physics. His first paper
was on Sulphide of Ammonium, his second on the Expansion of
Bismuth at the Freezing Point. His more important inquiries were
connected with the agglomeration of copper and the rarer metals by
the occlusion of hydrogen, and especially a series of experiments on

ANNUAL GENERAL MEETING. 353

the distribution of electricity in an electrolyte traversed by a
current. The curious and suggestive results of this investigation
appear in abstract in the Proceedings of the Royal Society for January
and June, 1881, and are most fully expounded in the second edition
of Mr. J. E. H. Gordon’s “ Treatise on Electricity and Magnetism.”

His most important research carried on in conjunction with
Dr. Gladstone is contained in a series of papers on the Copper-zinc
Couple, published principally in the Journal of the Society. He was
the first to observe the greatly enhanced chemical power of zinc when
covered with spongy copper. He was also the originator of another
long series of papers in our Journal on the Aluminium-iodine Reac-
tion, the last of which was read only the night before he was taken
ill. His discovery of these two new methods of acting on chemical
compounds, led to the discovery of many new substances, including
the aluminium alcohols and their products of decomposition.

In addition to these chemical inquiries, they worked together on
electrical matters —describing an air-battery, that is, one in which the
oxygen of the air took part; some experiments on thermal electrolysis ;
and, more particularly, the chemistry of Planté’s' and Faure’s secon-
dary batteries. The results of this investigation were collected
and published in a separate treatise. Mr. Tribe took out some patents
for a new modification of these secondary batteries.

As a teacher of science, Mr. Tribe was very successful. He had the
art of communicating his own enthusiasm to his laboratory students,
and his practical classes were well attended, although the boys had
to give up their Wednesday half-holidays for that purpose. Many of
these students have distinguished themselves since at the Universities
or elsewhere.

At these pursuits, Mr. Tribe worked earnestly and continuously,
being little known personally beyond his laboratories and his home.
His widow, and four surviving children, together with a small circle
of intimate friends, will however long remember the thorough
uprightness of his character, and the self-denying purpose of his life.

It was moved by Dr. Gladstone, and seconded by Professor
Thorpe, that the thanks of the meeting be given to the President for
his address, and that he be requested to allow it to be printed.
Referring to the publication of the catalogue, Dr. Gladstone said that
in looking through it he felt that it not only gave information as to
what was, but also what was not, in the library, and it might well serve
to lead many of us to make suggestions and to help to fill up gaps
from our own shelves. The resolution was unanimously accepted,
and was acknowledged by the President.

VOL. XLIX. 2B

354 ANNUAL GENERAL MEETING,

The Treasurer, Dr. Russell, then read his report, giving an expla-
nation of the balance sheet. He said that the funds of the Society
were in a satisfactory state, the income for the year having been
£3743 and the expenditure £3108, leaving a balance in hand of
£635, a sum that would make it possible to fund the life composi-
tions received during the year.

Mr. Carteighe moved that the thanks of the Society be tendered to
the Treasurer for his services during the past year; this was seconded
by Mr. Friswell.

*M. Crookes proposed a vote of thanks to the auditors, which was
seconded by Professor Clowes; Mr. Makins replied.

A vote of thanks to the Officers and Council having been proposed
by Mr. H. B. Dixon, seconded by Mr. Pickering, and acknowledged by
Dr. Armstrong,

Mr. Howard moved that the thanks of the meeting be given to
the Editor, Sub-editor, Abstractors and Librarian for their important
services to the Society during the year; this was seconded by Mr.
J. A. R. Newlands. Mr. Groves and Dr. Thorne replied.

Professor Humpidge and Dr. Teed were appointed scrutineers, and
a ballot having been taken, the following were declared to be elected
as Officers and Council for the ensuing year :—

President: Dr. Hugo Miiller.

Vice-Presidents who have filled the office of President: Sir F. A.
Abel; Dr. Warren De La Rue; Prof. Frankland; Dr. J. H. Gilbert ;
Dr. J. H. Gladstone; Professor A. W. Hofmann; Professor W.
Odling; Dr. W. H. Perkin; Sir Lyon Playfair; Sir H. E. Roscoe;
Professor A. W. Williamson.

Vice-Presidents: Mr. W. Crookes; Professor J. Dewar; Mr. David
Howard ; Professor G. D. Liveing ; Professor T. E. Thorpe ; Professor
W. A. Tilden.

Secretaries: Dr. H. E. Armstrong; Mr. J. Millar Thomson,
Foreign Secretary: Dr. F. R. Japp.
Treasurer; Dr. W. J. Russell.

Ordinary Members of Council; Messrs. H. T. Brown; Professor T.
Carnelley; M. Carteighe; Professor Frank Clowes; A. E. Fletcher;
R. J. Friswell; Professor R. Meldola; R.Messel; J. A. R. Newlands;
S. U. Pickering; Professor W. Ramsay; Thomas Stevenson.

TO Marcu 22, 1886.

Tur Terasvrnn txt ACCOUNT WITH THE CHEMICAL SocieTy FROM MARCH 23, 1885,

"9881 ‘NIFZ You
“NVISNAC] "}] KVHGNA \ ) "490.1100

9 1 oeg't|——

anaene JoinsveL jospueyur “
seeeereeesensens WUUBET 9B QOUBLET a
490} “Jua0 sod S¢ syi0 AM J0 puvog ueyodosjoyy ONEF JO Vseyoung
sjusuasuNgsig YseO 4430g 8,10INsBely, ma
0 0 QDR “tttttttetttteteeeteeeeeeees QR “gmD a
* giredoqy dod $¢ syiom jo prvog ueyyodose7
guaog avy 01 AIM 0 0 BBL rtttttsteeteeeeessteeees ynoag aunguaqed
0
8

“+ gasuadxg esnoy ANag Semirey ui0jsoM YWON pue uopuoT
wes s9NOg aso Jo sade Ay 0 000'F steeeeeeeesseseees § S1OSUOD “JUD 40d a4
an... Suruvat5 ° JaINSVALT JO spusy UT

sunysvy OL a See't seeeeeuseensereseensessesooes WUIBET 4B QIUBIUT
eeesecoses Surpimg eu) Survey ok a P
****""* SUUQUIYSAjoy SUPA ‘os anal

*sasuadagy asnoy

ee eeeeeeeeeenereseee oe seeeeeeeee “
sone teteeeeeeeeeees

Perrrrrr iit

MONOOMWNS

* syyviq pus ‘pry [eoa[D ‘sdureig 8 ees
ents suoyidiiosqns UO UOISSITIWIOD 8 wo
‘Ox ‘SPUlIG MUPUT AA “SudTUNG-sBs) 07 sure “es seasasene cosesecoesorers YOOIG OINjUIqGaCT
seeoee (guvok me Jepio ul snyeueddy ZuiqsmZunxy eq Suidoon Keaqrey U191890M YUMON pus UuOpUoT os
“ap ‘adeysog ‘k1au01g8Ig seeneseseeeeeees QOS “AOD
** xBj-aul00Uy pus Ajnqg ssnNoy peyqeyuy dad $€ S¥1OM JO pavog uvyodoijey bad
oti sesseeecovees BUMULIG SNOOUBT[IIS]]T “"** slosuog UO pPUEplAId
“+ guodeg JO 898] YIM Sspieg 4sog
sreerecseeeees SAMIDIOI JO 8PBIIsqy Juyuug
anaes £y0100g§ jesoy jO sdurpas.0i1g Zunquysiq
cvescccococococcocssoossoocososcoes saidog suowny Suguug

HPOOe eee e eee eee eee eneeeerereeeeeees

seosevassaoescorcconsacecee seeseseesessoor marae, sunnud

JO osuedxo SpIVMO} 4LIQIIDH “IG Wo. UOINGLWUOD

panes “* £98397] s¥iiq 9y} UO pusplAIg poullepoun uy

- ~-sqsdreuy oqng Jo Aja100g Ul0Ay -
sereeveres KIGSTULBYD JO 9INIIYSUT WOAT ”

verses KQQt00g peAoy JO sZurpso001g 10) suondiosqng

seseeeanssseenesesenvansseeeseessees YODUT DUB [BUINOL JO a[8G

aHCoOSCHOODrnow
HAAHM&ee- DODO
-——

“+ "oop ‘sappey AreiqrT
* SIOQUINN puv 819710] JeWeUy [BUOTIIPPY
SOUO Plo 04 SUOTIELA}[G PUL SasHOHOOY MON
PPrrTiT Titi seeeeeeeee * onZoleyep MOU 04 piezo
Yas Sad]A198 105 eul0u “iq pus uoIZULIeM “1X 0} WNLIeIOU0TT
PrPrrrrrrrirr rrr eeteeeeee s[BoIpOLlod 105 soxog pus Zuipuig
seeeecese@[BOIPOLlod PUB syood
quepueyy Aseiqry of
PPeeerieiic ii) uBLIBAqry jo Arvyeg
*havsqry £0 yunovp Uo sasusdzgy

oe weeeweeereee

ecmoocon o SO
cacota Oo °°

2
2
:
:
:
Z

Ae Ce eeeeeeeeeeeeeeseeeeeseeeeenes

aneenene 2881 405 suondisosqng z Z

“"""" s99q UOISSIUIPY GOT
Peer eee ereeeeeee suorisodui0g OT 2%

"988I ‘PUuze
yoy 07 ‘GEst ‘pigz youDyy wolf ‘sworjrsoduwo)
afvT pun ‘suondrwwsqng ‘saa uowsrupp Aq s)dravay

coocoooooeo
eocooocooceo

see eeeeeegeseeseeseeers

** pemnor jo uoNNquIsIG

* S10]98.19SQ VY 03 8904

cibabaateneing s0upa jo sosuaedxy [eyueprouy
“"10}IPA-qng aya jo Areypeg * I oze'1\—_—

*103IPYy 943 Jo Areleg |° noon *“1OINSVIIT, JO spury ur o

> ‘nie? g
pusnos fo zunoIp Vo sasuadry ~~ C881 “Pies Yue YUVg Je souvleg

OL ‘ESRI ‘22 HOUVW WONA ALAIOO§ IVOINAHD AHL AIIM INDOSSY Ni UANASVIANL aa =a

fe
=|
5
za
=|
a
ca
7,
fal
a)
nL
=
P
4
v4
<

‘NVISNOQ “Y WVHAONAM
‘SNINVIA ‘H ‘YD "OS8T “Y7bS YOTT (eesiitwniasens '
‘qooui00 punos pus poyIpny = epee yas 78 souveg
0 OoL‘s RP ER OSH ESTE SENS CORE REA I ES Se ee
*quoo aod $g syuoM JO pavog uvyjodoijsy
0 O000'T ** Y0}g “yu00 sed % Avmprey YsIWg YON
ae *sjassy

8 €& 86hF 8 & S86hF
cZL Re ee ee eee ee ae
8YA0M, JO pavog uvjzyodoajey_ uo puepiaig
c g 8g tenia ici
. yoojg Aumpwey YsMUG YON uo puspiaig

98T (9 0 cot **°*** Suudu0g ,suodeugg wosz uoyduaesqng ‘11 4me
eee eee ee eee eee eee eee ee OpvUL szUBLD “snp OL &I 602 00 99:60 C8 GESO SSO CONS 099) Be OS GORE "SSSI

‘esl ‘p's F

Tee eee eee eee eee eee ee yueq 48 oouvTegT
CHON STON EO DOS 054998 0409 50 Game S}UBLL *£renaqo qT

‘O88T “ANZZ HOUVIT OL ‘eget ‘augz HOUV]_, KOU ad
‘dNOJ HOUVASAY AHL HLIM INOONV NI ALALINOG ‘IVOINGHD AHL 40 AAAOSVAAT, ANY,

XXXIV.—The Relation of Benzalmalonie Acid to its Mononitro-
derivatives.

By Cartes M. Srvuart, M.A., Fellow of St. John’s College,
Cambridge.

In previous communications (Trans., 1883, 43, 403 ; 1885, 4'7, 155),
I have described benzalmalonic acid and its nitro-derivatives, and
remarked on the stability of the orthonitro-derivative as compared
with the corresponding para- and meta-derivatives, and with the un-
nitrated acid. It seemed that if any comparative measure of this
stability could be obtained, it would be a useful contribution to the
history of isomerism, moreover, the substances seemed well adapted
for this study.

When benzalmalonic acid is boiled with water, the following re-
actions take place :—

I. C,H;;CH: C(COOH), + H,O = O.H;COH + CH~COOH):.
Benzoic aldehyde. Malonic acid.
II. C,H,-CH : C(COOH), = C,H;CH : CH-COOH + CO,.

Cinnamic acid.

The three nitro-derivatives undergo decomposition in a similar
manner, for although I stated in my last communication that the
orthonitro-derivative was not decomposed on boiling, I find now that
a slight decomposition takes place. The amount of this decomposi-
tion was measured by the following method :—

A number of centigrams corresponding with the molecular weight
of the acid was boiled with 150 c.c. of water in a flask connected
with a reflux condenser, at the upper end of which was a calcium
chloride tube and potash apparatus for absorbing carbon dioxide as in
organic analysis. In the case of the unnitrated acid, agitation with
ether sufficed to remove benzaldehyde, cinnamic acid, and un-
altered benzalmalonic acid, leaving malonic acid in solution, which
was then determined by means of a seminormal solution of potash,
phenolphthalein being used as indicator, and a correction made for
the solubility of malonic acid in ether.

By this means, the carbon dioxide gave the amount of decomposi-
tion according to equation II, and the malonic acid that corresponding
to equation I. In the case of the nitrated acid, it was found better to
determine directly the amount of aldehyde produced; this was done
by neutralising with sodium carbonate, filtering off any precipitated
aldehyde, and agitating the filtrate with ether; the ethereal solution

358 STUART: THE RELATION OF BENZALMALONIC ACID

was then dried with calcium chloride and the ether distilled off.
The residue, together with the aldehyde collected by filtration, which
had in the meantime been allowed to drain on blotting paper, was dis-
solved in chloroform, and the solution filtered and evaporated in a
counterpoised dish, the residue in the dish being allowed to dry in a
vacuum over sulphuric acid. In this way remarkably concordant
results were obtained.

Benzalmalonic Acid.

KHO sol. CO,.

Boiled for ‘ I, 31:00 e.c. undetermined.

one hour. ) 1. 3122 ,, 0°0476 gram.

III. 30°72 ,, 0°0469 =,
Boiled for I. 21°05 ,, 0°0062_—=C«,,
15 minutes. | IT. 21-20 ,, 0°0055_—=C«,

A solution of malonic acid in water being made of approximately
the same strength, on agitation with the same amount of ether, it was
found that the ether dissolved } part of the malonic acid in solution,
z's of the number of c.c. of KHO are therefore to be added in each
experiment.

Paranitrobenzalmalonic acid.
Paranitrobenzaldehyde. CO..
Boiled for { I, 1:2120 gram. 0°6550 gram.
one hour. | II. 1:2495 __,, 00548,

Boiled ol I. 0°9293 __,, 00170 _ =sé—r,,

15 minutes. | If. 09458 __,, 070166,

Metanitrobenzalmalonic Acid.

Aldehyde. CO.

. 125 . 0458 gram.
Boiled = I, 1:2551 gram 0-0458 gram

one hour.

II. 1:2069 __,, 0°0634 =,
III. 12043 __, 0°0554
Boiled for 1 I. 09266 ,, 0:0280
15 minutes. | II. 09446, 0°0256

Orthonitrobenzalmalonic Acid.
Aldehyde.

I. (98 c.c. > KHO sol. used). 0°0252 gram.
Boiled for -

one hour. II. 0°3571 gram. 0°0287

IIT. 0°3692 __,, 00232
Boiled for ' I. 01314 _ ,, 00050
15 minutes. II. 01315 __,, 0 0055

TO ITS MONONITRO-DERIVATIVES. 359

Boiled for 1 hour. | Boiled for 15 min.
Decomposition Decomposition
according to according to

Equa- | Equa- Equa-
tion | tion | Total.] tion
I. II.

Paranitrobenzalmalonic acid ....

eS
a

” ”

Metanitrobenzalmalonic acid ........

” ”

3-1 ©
TS
— es
Ora Neko Ne SO
BAT HEAR BHO AO

” ”

Orthonitrobenzalmalonic acid ........

” ”

ie CO © © &

me > or

” ”

bo bo bo

Bromine Additive Products.

Another important difference between the orthonitro-acid on the
one side, and the unnitrated acid with its para- and meta-nitro-deriva-
tives on the other, is in the behaviour of these compounds when treated
with bromine, or hydrobromic acid. The combined bromine was
estimated by heating the substance with lime, dissolving in nitric
acid, adding to the filtered solution a known quantity of decinormal
solution of silver nitrate, and estimating the excess of silver by means
of decinormal thiocyanate solution.

Benzalmalonic Acid and Hydrobromie Acid.

Benzalmalonic acid was placed in coneentrated hydrobromic acid,
allowed to remain for two or three days, and then filtered through a
platinum cone; it was freed from excess of hydrobromic acid by
being placed under a bell-jar with solid potash. A white amorphous
powder was left, which on analysis gave the following results :—

0°2742 gram substance required 10°7 c.c. o AgNO; solution.
= 29°09 per cent. Br. Theory.... 29°30.

The substance, therefore, has the formula of phenyl-bromisosuccinic
acid, C,H;*CHBr-CH(COOH),.
lt cannot be purified, as when treated with water it dissolves,

360 STUART: THE RELATION OF BENZALMALONIC ACID

and in a few minutes a turbidity appears in the liquid which in-
creases until a mass of crystals of cinnamic acid is formed, whilst.
the solution is found to contain hydrobromic acid, the change
being expressed by the equation C,H;CHBr-CH(COOH), =
C,H,-CH : CH-COOH + CO,+ HBr. On treating this acid with
alcohol, it dissolves readily, and on allowing the alcohol to evaporate,
benzalmalonic acid is left, the alcohol having removed HBr.

Benzalmalonic Acid and Bromine.

Benzalmalonic acid was suspended in chloroform and an amount of
bromine equivalent to 1 mol. added; the liquid became slowly de-
colorised. The deposit was collected, washed with chloroform, and
analysed.

0°4235 gram required 22°6 in AgNO, solution.
{

= 42°62 per cent. Br. Theory.... 43°18.

It is, therefore, phenyldibromisosuccinic acid,
C.H;-CHBr-CBr(COOH),.
It melts at 96°, with evolution of hydrogen bromide, leaving a residue
which again becomes liquid with effervescence at 130—140°.
On treatment with water, it at once dissolves, but in a few minutes
the solution becomes turbid, and the a-bromocinnamic acid described
by Claisen, m. p. 131°, is deposited. The reaction being—

C.H;CHBr-CBr(COOH), = C,H;-CH: CBrrCOOH + CO, + HBr,

In experimenting with ethyl benzalmalonate, I was surprised to find
one carefully distilled specimen crystallised in large transparent
crystals, melting at 32°. Claisen mentions it as an oil, and I had not
previously observed the crystallisation.

0°471 gram gave 0°277 gram H,O and 1°1637 gram CQ).
Found. Theory.
67°74
6°45
This ethyl salt was treated with hydrobromic acid, but the product
obtained was an oily substance which could not be purified suffici-
ently for analysis.

Metanitrobenzalmalonic Acid and Hydrobromic Acid.

This acid behaves in exactly the same way as the unnitrated acid,
a white amorphous residue is obtained, of which on analysis—

TO ITS MONONITRO-DERIVATIVES,

0:2722 gram required 8°45 a Ag.
Found.. 2483 per cent. Br. Theory .

It is, therefore, metanitrophenyl-f8-bromisosuccinic acid.

It is decomposed by water with the formation of metanitro-
cinnamic acid, and by alcohol with the formation of metanitrobenzal-
malonic acid in the same way as the derivative of the unnitrated acid

described above.

Metanitrobenzalmalonic Acid and Bromine.

Metanitrobenzalmalonic acid was suspended in chloroform, and an
amount of bromine equivalent to 1 mol. added, the solution became
decolorised, and the residue after being washed with chloroform was
analysed.

0°4714 gram required 23°9 c.c. 0 AgNO, solution.

= 40°56 per cent. Br. Theory .. 40°30

It is, therefore, metanitrophenyldibromisosuccinic acid. When
heated, it gives off hydrogen bromide at 120—140°, and melts with
effervescence at 195°. It dissolves in water, and the solution quickly
becomes turbid and deposits white crystals melting at 212°, of which
on analysis—

0°3674 gram required 13°6 ip AgNO, solution.

Theory for metanitro-a-
bromocinnamic acid,

= 29°85 per cent. 29°41

It is, therefore, metanitro-2-bromocinnamic acid.

Ethyl metanitrobenzalmalonate, m. p. 73°, dissolved in concen-
trated hydrobromic acid, and after a time the solution deposited
crystals which were collected and recrystallised from bisulphide of
carbon ; on analysis—

0°376 required 10-4 c.c. 7 AgNO; solution.

= 22°12 per cent. Br. Theory.. 21°39 per cent.

It has the formula C,H,(NO,)-CHBr-CH(COOC.H,)., and is tolerably
stable, as it can be melted (m. p. 88°) or crystallised without under-
going alteration, and is not decomposed by water.

362 STUART: THE RELATION OF BENZALMALONIO ACID

Paranitrobenzalmalonic Acid and Hydrobromic Acid.

When paranitrobenzalmalonic acid is placed in concentrated hydro-
bromic acid, a residue is obtained which can be freed from water and
hydrobromic acid by being placed under a bell-jar with potash. On
analysis—

0°8793 gram gave 0°5261 AgBr.
= 25°46 per cent. Br. Theory .. 25°15

It dissolves in water, and the solution rapidly becomes turbid and
deposits paranitrocinnamic acid, m. p. 286°; with alcohol it gives
para-nitrobenzalmalonic acid.

Paranitrobenzalmalonic Acid and Bromine.

If paranitrobenzalmalonic acid is suspended in chloroform, and an
amount of bromine equivalent to one molecule added, the colour of
the solution becomes gradually paler, but is not entirely decolorised.
The deposit, after being collected and washed with chloroform,
appears to be a mixture of the paranitrobenzalmalonic acid with its
dibrominated derivative, as the percentage of bromine found was
always low, and the substance was only partially dissolved by water,
with subsequent deposition of crystals. Paranitrobenzalmalonic acid
is insoluble in chloroform, and is only partly converted into the
bromo-additive product by the above method. The action of bromine
itself on the acid was then tried. A specimen of paranitrobenzal-
malonic acid was placed in bromine in a stoppered cylinder, which
was frequently shaken, and at the end of a month chloroform was
added, and the mixture filtered. The residue was then washed with
chloroform and allowed to dry; it gives off hydrogen bromide at
90—100°, and the residue melts with effervescence at 188°; the
residue left melts at 250—260°.

On analysis—

0°5256 gram required 26 8 c.c. in AgNO, solution.
= 40°79. Theory.... 40°30

it is therefore paranitrophenyldibromisosuccinic acid. On treating it
with water, it dissolves, and the solution deposits needles melting at
2u8°. Of these—

0°1402 gram required 5:17 c.c. i AgNO, solution.
= 29°50. Theory .. 29°41

TO ITS MONONITRO-DERIVATIVES. 363

The substance, therefore, is the paranitro-a-bromocinnamic acid,
described by Muller (Annalen, 212, 122) as the f-acid.

Ethyl paranitrobenzalmalonate dissolves in concentrated hydro-
bromic acid, and the solution deposits crystals. These, after being
washed with water and crystallised from bisulphide of carbon, were
analysed :—

0°4208 gram required 11 c.c. va AgNO, solution.

= 20°95 per cent. Br. Theory.. 21:39

It is therefore ethyl paranitrophenyl-8-bromisosuccinate. It melts
at 89°, and is stable, as it can be crystallised from alcohol or bisulphide
of carbon, or washed with water without alteration.

Addition-products of Orihonitrobenzalmalonic Acid.

Orthonitrobenzalmalonic acid behaved quite differently from the
compounds above described. If placed in concentrated hydrobromic
acid, a yellow substance was deposited, and a further quantity of the
same compound was obtained on diluting the liquid. It was not decom-
posed or dissolved by water, and is insoluble in benzene, ether, chloro-
form, and bisulphide of carbon. It dissolves readily in alcohol, but
could not be obtained crystallised. It melts without effervescence at
about 227°. Two analyses gave—

41°69 per cent. Br.
474 yw

Theory for additive compound with HBr similar to the others,
25°15 per cent.

A similar compound is obtained when orthonitrobenzalmalonic acid
is treated with chloroform and bromine.

Ethyl orthonitrobenzalmalonate dissolves in concentrated hydro-
bromic acid, and the solution deposits crystals, which after being
collected, washed with water, and crystallised from bisulphide of
carbon, were submitted to analysis.

0°4729 gram required 12°77 c.c. i AgNO; solution.
= 21°60 per cent. Br. Theory.. 21°39

It is orthonitrophenyl-8-bromisosuccinate. It melts at 68°, and can
be crystallised from the usual solvents; it is not decomposed by water.

The following is a list of the bromo-additive compounds de-
scribed :—

364 STUART: THE RELATION OF BENZALMALONIC ACID,

| ( Decomposed by water,
Benzalmalonic acid + HBr...... Additive product. forming cinnamic
p-Nitrobenzalmalonic acid + HBr | Do. and nitrocinnamic

acids.
m-Nitrobenzalmalonic acid + HBr Do. Decomposed by alco-

hol, forming ber zal-
|. malonie acid.
o-Nitrobenzalmalonic acid + HBr | Condensation product. | Not decomposed by

water or alcohol.

| { Decomposed by water,
forming a-bromo-
Benzalmalonic acid + Br, |) Additive product. | cinnamic acids, give
p-Nitromalonic acid + Br, } Do, + off HBr at 100—
m-Nitromalonic acid + Br,......| Do. 130°, leaving a re-
| sidue melting with
| evolution of COs.
| Condensation product. | Not decomposed by

o-Nitromalonic acid + Br,
water or ulcohol.

Ethyl - p - nitrobenzalmalonate + |
BAM cece sugecaeccsnceees Additive product.
Ethyl -m- nitrobenzalmalonate +

} Stable compounds not
Do \ decomposed by water,
; crystallised from alco-

Do. hol or CS.

These experiments confirm a statement of Miiller (Annalen, 212,
122), that the ethyl salt of an unsaturated acid exhibits a greater
tendency to take up bromine and pass into a saturated compound than
.the acid itself.

I believe that these experiments may contribute something to our
ideas as to the structure of benzene. The two symbols which at
present claim to represent the structure of benzene are Kekulé’s
hexagon and Ladenburg’s prism. In the former, the carbon-atom in
the ortho-position is directly connected with the first carbon-atom,
whilst those in the para- and meta-positions are indirectly connected
with it ; in the latter the carbon-atoms in the para- and meta-positions
are directly counected with the first carbon-atom, whilst the carbon-
atom in the ortho-position is indirectly connected with it. In K6érner’s
modified hexagon, which is, however, now rejected as showing only two
di-derivatives, the carbon-atoms in the ortho- and para-positions are
directly connected with the first carbon-atom, and that in the meta-
position is indirectly connected with it. Now assuming that a nitro-
group attached to one carbon-atom is more likely to modify the
stability of a hydrocarbon chain attached to a carbon-atom which is
directly connected with it, more so than if that carbon-atom is
indirectly connected with it, we should expect to trace out analogies
in the following way :—

According to Ladenburg’s prism, the para- and meta-nitro-deriva-

STUART: CINNAMIC AND SALICYLIC ALDEHYDES. 355

tives would behave similarly on the one hand, whilst the ortho-deriva-
tive and the unnitrated acid would behave similarly on the other; ir
the influence of the nitro-group be protective, the para- and meta.
derivatives should be more stable.

According to Kekulé’s hexagon, the para- and meta-derivatives
and the unnitrated acid would behave similarly on the one hand, and
the ortho-derivative would behave differently on the other; if the
influence of the nitro-group be protective, the ortho-derivative should
be more stable than the other three.

According to Kérner’s modified hexagon, the ortho- and para-deri-
vatives should behave similarly on the one hand, and the meta-deriva-
tive and the unnitrated acid should behave similarly on the other; if
the influence of the nitro-group be protective, the ortho- and the para-
derivatives should be more stable.

We see that of these three hypotheses, the second is the one with
which the results obtained are in accord, and I would suggest this as
an argument in favour of Kekulé’s hexagonal formula.

The words of R. Meyer (Ber., 15, 1826) in defence of the prism
formula against Kirner’s modified hexagon, are worth quoting in this
connection. He says, “It is assumed that the actual relatiouship of the
benzene compounds points rather to analogies between the ortho- and
para-derivatives. I do not believe that such an analogy can be
recognised.”

It was now thought that the action of other aldehydes on malonic
acid might also be utilised, with the view of ascertaining if any
alteration of stability was produced by altering the side hydrocarbon
chain, and I was surprised to find that a derivative of cinnamic alde-
hyde, containing the double bond twice over, was more stable than
auy of the compounds mentioned in this paper.

The results obtained are given in the succeeding paper.

XXXV.—<Action of Cinnamic and Salicylic Aldehydes on Malonic
Acid.

By Cuartes M. Sruart, M.A., Fellow of St. John’s College,
Cambridge.

A mixtTuRrE of equal parts of cinnamic aldehyde and malonic acid with
half its weight of glacial acetic acid was heated at 100° for six hours ;
the crystalline product, after being collected and washed successively
with glacial acetic acid, water, and chloroform, presented the appeur-
ance of yellow needles. On analysis—

STUART: CINNAMIC AND SALICYLIC ALDEHYDES,

0°3639 gram gave 0°8829 gram CO, and 0°1446 H,0.
4°41 per cent. Theory...... 459
66°05

It is therefore phenylbutindicarboxylic acid,
C.H,CH : CH:CH : C(COOH),.

It melts at 208° with effervescence, carbon dioxide being evolved.
The residue had no constant melting point, but if a larger quantity
was heated at 210° till it ceased to give off carbon dioxide, and the
product crystallised from benzene, pure cinnamenylacrylic acid, melt-
ing at 165° (Perkin, Chem. Soc. J., 31, 403), is obtained; the cinna-
menylacrylic acid is partially decomposed by heating above 200°.

Phenylbutindicarboxylic acid, unlike benzalmalonic acid, can be
boiled with water, without decomposition.

I now tried salicylaldehyde, with the view of ascertaining the
influence of a hydroxyl-group in the ortho-position.

A mixture of salicylaldehyde, malonic acid, and glacial acetic acid,
on being heated for some time at 100° and allowed to cool, became
semi-solid from the formation of crystals. These were drained from
acetic acid and recrystallised from hot water, when white needles,
melting at 187° without effervescence, were obtained.

On analysis—

I. 0°4347 gram gave 1:0056 gram CO, and 0°1288 gram H,0.
II. 0°3667 - 0°8489 is 0°1091 "

III. 0°3521 - 0°8131 - 0°1016 -
corresponding to—

I. II. III. Theory. CyoH,O,
63°09 63°13 62°97 63°15

3°29 3°30 3°21 3°16

The substance treated with dilute ammonia dissolved, and the
solution gave a white precipitate with silver nitrate, of which on
analysis—

0°6045 gram gave 0°2211 gram Ag,

corresponding to 36°57. Theory for C,H;O,Ag.... 36°36.

The barium salt was obtained as a white precipitate on adding
barium chloride to the same solution.

0°7117 gram dried at 130° gave 0°3229 gram BaSQ,.
= 26°67 per cent. Ba. Theory (C\H;0,).Ba.. 26°60.

This acid is therefore a monobasic acid of the composition

o—co
C.H,
\ ;
CH: C-COOH,

RIDEAL: ACTION OF AMMONIA ON CHROMYL DICHLORIDE. 367

and may be called coumarincarboxylic acid, It is not decomposed by
boiling with water, or on melting, but on heating it in a distilling
flask above its melting point it evolves carbon dioxide, and at 290° a
substance distils over and solidifies in the receiver. When recrystal-
lised, this melts at 67°, and its solution in sodium hydrate gives the
characteristic reactions of coumarin with metallic salts. The thermo-
meter then rises rapidly above the boiling point of mercury, and a small
quantity of the unchanged coumarincarboxylic acid distils over. An
attempt to produce a bibasic hydroxy-acid corresponding to coumaric
acid, by heating coumarincarboxylic acid with sodium ethylate,
according to Fittig and Ebert’s method (Annalen, 226, 351), led to
no result.

I am at present occupied with the production of benzalmalonic acid
containing some negative element, such as chlorine, in the ortho-
position, and hope soon to have the honour of laying my results before
the Society.

XXXVI.—Note on the Action of Ammonia on Chromyl Dichloride.

By Samvet Ripeat, B.Sc., University College, London.

Wuen the vapour of chromyl dichloride, CrO,Cl,, is passed into a
vessel containing dry ammonia gas, it takes fire and burns with a
white flame, giving, at first, white fumes, and finally a greenish-
brown solid. This reaction has been noticed and described by
various observers. Liebig thought that the brown compound formed
in this manner, and the black compound obtained when chromic
chloride, CrCl, is substituted for the chromyl dichloride, consisted
of metallic chromium (Pogg. Ann., 21, 359). Schrétter repeated the
experiment with the chromic chloride and analysed the product,
which he found to be a nitride of chromium having the formula
Cr;N,. He did not, however, examine the product of the reaction
with chromyl dichloride, but inferred that a compound of similar
constitution was formed. He also noticed that this chromium nitride,
when heated at 150—200° in a stream of oxygen, took fire and
burnt with a red light, evolving nitrogen and oxides of nitrogen.
(Annalen, 37, 145). To determine the correctness of Schrétter’s in-
ference, a small quantity of the substance obtained from chromyl
dichloride was heated in a piece of hard glass tubing and a current
of dry oxygen passed through it, but neither nitrogen nor oxides of
nitrogen were evolved, and the substance did not glow. It was,

368 RIDEAL: ACTION OF AMMONIA ON CHROMYL DICHLORIDE.

therefore, evident that the compound obtained from chromy] dichloride
differed from that derived from chromic chloride.

Thomson also examined the reaction between chromyl dichloride
and ammonia, and found that the dark brown solid, when digested
with water, left a green powder which was insoluble in nitric acid,
and contained chlorine. He also noticed that ammonic chloride was
one of the products of the reaction (Phil. Mag., 1827, 452). Persoz,
in a paper on the combination of gaseous ammonia with metallic
chlorides, has shown that chromyl dichloride combines with am-
monia in the ratio of 79°101 of the former to 20°898 of the latter;
these numbers give 3:8 as the ratio of the reacting molecules.

The present note describes some experiments which were under-
taken to correlate and complete these statements. As already men-
tioned, white fumes are conspicuous at the beginning of the reaction,
and when the experiment was conducted in a long tube filled with
dry gaseous ammonia, it was noticed that the product was not
homogeneous, but consisted of two substances, a volatile white solid,
and a non-volatile greenish-brown compound; a current of dry am-
monia was passed through the tube for some time after the reaction,
so as to ensure the complete decomposition of the chromy] dichloride.
The product was then digested with water, in which it was partially
soluble; the solution was yellow in colour, contained ammonium
chloride, chromic acid, and a substance which could be precipitated
as a brown powder by ammonia. This brown precipitate, when
boiled with potash, was converted into green chromic hydrate, and
the filtrate contained a chromate; it was therefore the brown oxide
of chromium, Cr,03,CrO;. The insoluble residue dissolved slowly in
hydrochloric acid, and appeared to consist only of chromic oxide.
After being well washed and boiled for some time with water, it con-
tained no chlorine.

It being probable from these results, that some of the ammonia
was decomposed with liberation of nitrogen, this gas was looked
for among the products. A weighed quantity of the chromy] di-
chloride in a thin glass bulb was placed in a large and strong glass
bottle connected with an apparatus for delivering a rapid current of
dry ammonia, and with a delivery tube dipping under mercury.
When all the air had been expelled from the apparatus the glass bulb
was broken and the gases evolved collected over water. Nitrogen
was given off in the proportion of 1 mol. to 3 mols. of chromy] dichlo-

ride. In two experiments, the following numbers were obtained :—

Volume of nitrogen.

Weight of r ~
chrowyl dichloride. Found. Theory.
2°0515 grams 96°4 c.c. 98°7

0°5835_—i, 28°6 ,, 28°05

Coord ss 8s ow Kk,

VELEY: SOME SULPHUR COMPOUNDS OF BARIUM. 369

The solid products obtained in these experiments were also collected
and analysed. Any uncombined ammonia which might have been
condensed on the solid product was removed by passing a current of
dry air over it, and after weighing the whole, the amount of am-
monia present in a weighed portion was determined. It yielded
22°1 per cent. of combined ammonia, which agrees with the per-
centage of ammonia (21°9) calculated on the assumption that 8 mols.
react with 3 mols. of chromy! dichloride in the following manner :—

3CrO,.Cl, + 8NH,; = 6NH,Cl + CrO;,Cr.03.

The amount of chromic acid in the product was found to be
21:5 per cent. (theory 21°7), and after ignition, the percentage of
chromic oxide left was 43°7 (theory 42°7).

The cbromy! dichloride, then, in this reaction, appears simply to
break up into the dioxide and chlorine, and the latter is then able to
react with the ammonia inthe ordinary manner. Chromy] dichloride,
when heated alone to dull redness, splits up into chromic oxide, oxygen,
and chlorine (Woéhler, Annalen, 60, 203), and when heated in a sealed
tube loses two-thirds of its chlorine, forming (CrO,);Cl., proving that
the chlorine is only loosely combined. With carbamide, Sell has
shown that a compound is formed containing no chlorine (Proc. Roy.
Soe., 33, 267).

The behaviour of chromyl dichloride towards the amines is now

being examined.

XXXVII.—Some Sulphur Compounds of Barium.

By V. H. Veter, M.A., of the Laboratory, Christ Church, Oxford.

Historical.

Barium Monosulphide.— Berzelius, in his monograph on the sul-
phides of the alkalis and alkaline earths (Schweigger’s Journal, 34, 12),
describes the preparation of barium sulphide by the well-known
method of subjecting a mixture of barium sulphate and coal to a red
heat, lixiviating the mixture, and concentrating by evaporation. A
purer product (Lehrbuch, 2, 137) was obtained by heating the oxide
or carbonate in an atmosphere of carbon disulphide, a method found
by Schéne to give quantitative results.

Barium Hydroxyhydrosulphide.—The composition of the crystals
obtained by lixiviating with water the washed mass of barium sul-

VOL. XLIX. 2c

370 VELEY : SOME SULPHUR CUMPOUNDS OF BARIUM,

phate and coal, and cooling or evaporating the solution, was examined
by Rése (Pogg. Ann., 55, 415), who noted the variation in com-
position and appearance of the crystals with the conditions of pre-
paration, such as temperature, time of contact of crystals with solution,
quantity of water used. By treating barium trisulphide with water
and evaporating in a vacuum, Schéne (Pogg. Ann., 112, 193) obtained
successively crystals of barium hydroxyhydrosulphide, BaSH-OH,5H,0,
the hydrosulphide (vide infra), and a tetrasulphide with variable
quantities of water of crystallisation.

Barium Hydrosulphide.—Berzelius (Pogg. Ann., 6, 441) describes
the preparation of the crystalline hydrosulphide by saturating a solu-
tion of the oxide in water with hydrogen sulphide, adding alcohol to
the solution, filtering off the sulphur and oxidised products, and
cooling to —10°, when groups of four-sided prisms separated; and
secondly, by evaporating the solution in a vacuum, when opaque, pro-
bably efficresced, acicular crystals were obtained. No analytical results
are given, but the quantitative test was applied of the precipitation
of manganese sulphide and evolution of hydrogen sulphide on addi-
tion of a solution of a manganese salt to a solution of the crystals.

The crystals obtained by Schéne gave on analysis results corre-
sponding to a formula Ba,S,,25H,0, but it is admitted that the crystals,
which were of a pale-yellow colour, were slightly impure. The
crystals lose water at 100—110°, at 230° evolve hydrogen sulphide,
and are not completely decomposed into the monosulphide and
hydrogen sulphide even at 280—300°.

The sulphides of barium have been examined from a thermochemical
point of view by Sabatier (Ann. Phys. Chem. [5], 22, 598).

Barium Thiocarbonate.—Berzelius (Pogg. Ann., 6, 244) describes
this salt as a yellow crystalline substance, sparingly soluble in water.

The preparation of anhydrous barium monosulphide, crystalline
barium hydrosulphide, and the conditions of formation of barium
thiocarbonate, form the principal subjects of the present commu-
nication.

Preparation of Barium Monosulphide, BaS.

The method adopted for the preparation of this substance con-
sisted in passing hydrogen sulphide over a known weight of barium
hydroxide, collecting the water evolved, and weighing the resultant
barium sulphide.

For this purpose, it was thought especially desirable to obtain the
hydroxide in a state of purity, which was effected by the following
method :—Recrystallised barium chloride was dissolved in water,
aud a few drops of ammonia and ammonium sulphide added to

VELEY: SOME SULPHUR COMPOUNDS OF BARIUM. 371

the solution. The small quantity of sulphide of iron precipitated
was removed by filtration, the solution was then evaporated to dry-
ness, and the barinm chloride rubbed up frequently with absolute
alcohol to dissolve out any strontium or calcium chloride, with
which the substance is likely to be contaminated. In order to
convert the chloride into the hydroxide, the method proposed by
Mohr and found successful by Bloxam was used, which consists in
outline of boiling the barium chloride with sodium hydrate solution
(sp. gr. 1:00—1°15). The scdium hydrate used was purified by
solution in alcohol, and in order to eliminate as far as possible the
presence of carbonic anhydride and consequent formation of barium
carbonate, the process was conducted in vessels filled with hydrogen.
The crystals of Ba(OH).,8H,O were recrystallised several times,
always in an atmosphere of hydrogen, finally thrown on a filter, and
washed with cold, recently boiled water, until the wash-water gave
no precipitate with silver nitrate. The crystals thus obtained con-
tained no carbonate. They were then dried over a mixture of potas-
sium hydrate and calcium chloride, and before treatment with
hydrogen sulphide, were heated in a current of hydrogen at about 80°,
until their weight was constant. The composition of the substance
was found to be Ba(OH),,H,0, as noticed by previous observers.

Found. Theory.
Ratio of BMOH):H:O | 1997 1-281
BaO

It is generally stated in the text-books that crystalline barium
hydrate melts in its water of crystallisation at 100°. Experiments,
however, showed that the temperature is 83—85°; mean of observa-
tions made (i) by enclosing sone of the substance in a capillary tube
filled with hydrogen, and heating in a bath of glycerol; (ii) by immersing
a thermometer into a considerable quantity of the substance heated
in a small test-tube, through which a continuous current of hydrogen
was passed.

The apparatus used for the conversion of the hydroxide into the
sulphide was identical with that described in my paper on some
sulphur compounds of calcium (Trans., 1885, 481).

Methods of Analysis used. — Sulphur was determined by Carius’
method. Barium was estimated by evaporating the substance with
dilute sulphuric acid in a porcelain crucible and weighing the sulphate
formed, and water by heating the substance in hydrogen, or preferably
in carbonic anhydride, owing to the retention of water by barium
hydroxide.

Preliminary experiments indicated that the sulphur compound ob-
tained on heating the monohydrate of barium hydroxide in hydrogen

2c2

372 VELEY : SOME SULPHUR COMPOUNDS OF BARIUM.
sulphide retained, at temperatures above 160°, some hydrogen sul-
phide, which is given off again at a higher temperature ; it is only by
prolonged heating at 200° that the conversion of the hydroxide into
the monosulphide can be completed.

The change was investigated both by synthetical and analytical

methods.

Synthesis.

Conditions of experiment.

sitcan I Weight
| —_ ra! Time. of |
— material. |
|

|
Differ- |

ence,

Water
collected.

Hydrate of barium hydroxide

grams.
2°1484 |

grams. grams.

Heated in a current of hydro-
gen sulphide
Heated in a current of hydro-

| 120-125 |
| 130—135 |

mr DO DO bo bo

”

+0 °3539

+0°1775

+0°0278

+0°058
nil.

Barium sulphide .......

Analysis of Resultant Substance.

Percentage
found.

Sulphur as barium sulphate, by

Carius’ method

Hydrate of barium

Ba(OH).,H,O

Barium sulphide, BaS obtained

W ater collected

Synthesis.
Theory.
hydroxide,
2°1484
19234
0°6147

Theory for
BaS.
18°94
81:06

100°00

Experiment.
2°1484
1:9338
0°6172

A similar series of experiments were conducted, but owing to an
accident, the determination of water was lost.

VELEY : SOME SULPHUR COMPOUNDS OF BARIUM. 373

Series II.

Conditions of experiment.
material.

grams. | grams.
Hydrate of barium hydroxide 1°838 _

Heated in a current of hydrogen sul- { 2 HH yy a —
IE caciansansnaxhessunhonsda ee 100_110}| 19225 | +0-0878
Heated in a current of hydrogen ....| 14 hrs. | 155—160 | 1°772 —0°1535
” eee 3 brs. | 200—210 | 1°6597 | —0°1023

‘ ' aes | 210—215 | 1°66 —0 0097

” cece 1°6505 | —0:0098

% , sees 1-630 —0°0005

1°650 _

sarium sulphide

Analysis of Resultant Substance.
Percentage Theory for
found. BaS.

Sulphur 18°94
81:06

100°00

Synthesis.
Theory. Experiment,
Hydrate of barium hydroxide.. 1°838 1-838
Barium sulphide 1°650

Thus both by synthetical and analytical processes it is shown that
the hydrate of barium hydroxide is converted into barium sulphide in
accordance with the equation—

Ba(OH),,H,O + H,S = BaS + 3H,0.

Anhydrous barium sulphide, which apparently has not previously
been described, is a white amorphous powder, soluble in water, but
perhaps only after decomposition into the hydroxide and _ hydro-
sulphide. On exposure to air, it changes to a pale yellow, and then
to an orange colour, owing to the formation of oxidised products. On
addition of concentrated nitric acid, its heat of oxidation is so great
that it becomes vividly incandescent. If it contains a trace of iron,
the sulphide is of a pale pink colour, but with a larger excess it has
the same colour as ordinary fresh lime.

VELEY : SOME SULPHUR COMPOUNDS OF BARIUM.

Liziviation of Anhydrous Barium Sulphide with Water.

It is well known that when anhydrous barium sulphate, obtained by
the reduction of the sulphate with coal or other carbonaceous matter,
is heated with an excess of water, it is decomposed completely into
the hydrosulphide, which dissolves, and the hydroxide, which in part
dissolves, in part crystallises out. But if a smaller quantity of water
be added, six-sided tables of barium hydroxyhydrosulphide separate,
admixed with variable proportions of the hydroxide. As the writer
had occasion to prepare a considerable quantity of the former sub-
stance, and as no details have been given regarding the time of
boiling, the quantity of water used for lixiviation of the material, &c.,
a series of comparative experiments were made to determine the best
conditions. In each experiment the following quantities were
used :—

150 grams.
50s,
10 9

These were well mixed and heated in a crucible to a red heat for
one hour. In order to equalise the conditions as far as possible, the
materials were always taken from the same sample, the same crucible
and lid were used, and the roasting performed at the same time of
day, when the pressure of gas supplied was the same. After cooling,
the roasted mass was placed in a flask, and variable quantities of
boiling water poured upon it, the whole being boiled for four
minutes. It was then rapidly filtered and cooled, the vessel being
well agitated to promote the separation of crystals. The crystalline
deposit was allowed to stand until the supernatant liquor was
perfectly cool, collected, washed once with alcohol to remove the
adherent solution of the hydrosulphide, and dried between folds of
blotting-paper. Though the results are of course not rigidly com-
parable, yet they are sufficient to show both the variation of yield,
and of the proportion of barium sulphide contained therein, with the
proportion of water used for the crystallisation.

Percentage of

Volume Weight of barium sulphide
of water. crystals. in crystals.

250 42 grams 43°1
300 58 Cs, 41°13
300 56 C«s 42°48
350 39 COC, 28°90
400 10 (barium hydroxide) nil.
450 nil, nil.
These crystals owing to their ready oxidisability do not offer a con-

VELEY : SOME SULPHUR COMPOUNDS OF BARIUM. 375

venient material for the preparation either of the anhydrous sulphide
or of the hydrosulphide.

Barium Hydrosulphide, Ba(SH),,4H,0.

To prepare this substance, the method described by Divers (Trans.,
1884, 271) for the corresponding calcium compound was used with
suitable moditications. In the preparation tube, a solution of the
crystalline barium hydroxide, saturated at 100°, was made in a con-
tinuous current of hydrogen, which was then replaced by hydrogen
sulphide purified from oxygen and dried. A considerable develop-
ment of heat accompanied the absorption of the gas. To obtain a
saturated solution, the hydroxide dried at 100°, or preferably the
anhydrous oxide was added, the solution being kept at a temperature
of 60—70°, while a continuous current of hydrogen sulphide was
passed in. This process requires several days for completion. Ulti-
mately crystals separate from the solution when cold. To obtain
these free from undissolved impurities such as silica, ferric sulphide,
and unaltered oxide, the solution was again warmed to 40°, and the
supernatant liquid quickly decanted off into a small stoppered bottle,
filled with hydrogen sulphide. The liquid was then cooled, and finally
placed in a freezing mixture. Tufts of hard, interlaced acicular
crystals separate, which were allowed to grow for some days in the
saturated solution. The mother-liquor was drained off, the crystals
quickly detached, and dried between folds of blotting-paper, so
arranged as to allow of the least possible access of air. When these
were dissolved in water and a solution of a manganese salt added,
manganese sulphide was precipitated with evolution of hydrogen
sulphide ; the crystals were insoluble in alcohol. These qualitative
tests indicate that the crystals consist of barium hydrosulphide. Their
analysis was effected as follows: a weighed quantity was treated
according to Carius’ method, and from the barium sulphate at first
formed was determined the barium and sulphur existing in the form
of monosulphide. Ou filtering and precipitating the filtrate with
barium chloride, the sulphur in excess of that required for the mono-
salphide was determined. If the barium hydrosulphide were
perfectly pure these two quantities of barium sulphate should be the
same.

Analysis of First Sample.
Percentage

Barium. Sulphur.
0°4024 gram gave (i) 0°3334 gram BaSQ,.. 48°7 11°07 (i)
(ii) 0°3158 - o_o = 10°78 (ii)

48°7 22°15

” ”?

376 VELEY : SOME SULPHUR COMPOUNDS OF BARIUM.

Atomic ratio of barium to sulphur = _

Analysis of Second Sample.

Percentage
ie — a.
Barium. Sulphur.
04524 gram gave (i) 0°3861 gram BaSO,.. 50°14
(ii) 0°3852 ies

50°14

9 ”

Atomic ratio of barium to sulphur 1 : 1°983.

0°2634 gram heated in a current of carbonic anhydride gave 0°068°3
gram water, or 25°92 per cent.

These last results point to a formula, Ba(SH),,4H,0, for the

crystals.
Percentage Theory for
results. Ba(SH)>,4H,O. Found.
Barium 49°81 50°14
Sulphur (i) .... ng >: 1 11°71 1 oc.
23°2 23°40
» (ii).... 11°64 . 11°69 J
25°92

By difference.. "54
100-00

The crystals when heated in a sealed capillary tube filled with
hydrogen exhibited no signs of liquefaction in their water of crystal-
lisation, but were simply dehydrated, a result in accordance with the
observations of Berzelius.

Temperature of Decomposition of Barium Hydrosulphide.

The following experiment was made to test the accuracy of
Schéne’s statement that the hydrosulphide does not evolve hydrogen
sulphide at temperatures below 260°. A weighed quantity of the
hydrosulphide in a porcelain boat, enclosed within a piece of com-
bustion tubing, was heated in an air-bath in acurrent of hydrogen, the
gas being purified successively by chromic acid, sulphuric acid, potash,
damp oxide of iron, and finally dried with caicium chloride. At the
outlet of the (J tube containing the damp oxide, a small slip of paper
moistened with lead acetate was placed. The gas after passing over
the hydrosulphide was passed into a Liebiy’s bulb apparatus, filled
with the lead syrup used for the estimation of sulphur compounds in
coal-gas, and placed upon a sheet of white paper. The air-bath was

VELEY: SOME SULPHUR COMPOUNDS OF BARIUM. 377

provided with a thermometer enclosed within a similar piece of
combustion tubing in order to make its surroundings and temperature
identical with that of the hydrosulphide. A blank experiment showed
that no coloration was given to the lead solution on passing the
hydrogen through the apparatus for one hour. The crystals were
then introduced, and the hydrogen passed for a short time, before
connecting the bulb apparatus, in order to blow off any hydrogen
sulphide liberated by the temporary exposure of the crystals to the
carbonic anhydride of the atmosphere. At a temperature of 18°,
no decomposition was observed, the lead solution remaining unaltered
for half an hour. At a temperature of 50°, the lead solution gradually
darkened, thus showing an incipient decomposition ; at 55°, the colora-
tion became more marked, while at 60° the lead sulphide was
completely precipitated. This decomposition was concomitant with
the elimination of water, as at 55° a distinct dew was observed in the
cooler portion of the combustion tubing, which increased considerably
at 60°. But when the substance is completely dehydrated its stability
is remarkably increased, the hydrogen sulphide being retained with
great persistence. The results in the following table illustrate the
statements above.

Conditions of experiment.

| g
eee -

Heated in a current of hydrogen| 3 hrs. 100—105° : 0° 0836
- 0-009
112—115 0 ‘0036
145—150 ;
170—175 0-001
195—200 D° 0-002
235—240 : nil.
260-265 : 0-001
Red heat . | 00198

Lose per comt. 22.002 cccces cece | 37°85
Loss required by theory for con-

version of Ba(SH).,4H,O tu
38°18

Thus the hydrogen sulphide is only completely eliminated at a red
heat ; but at a lower temperature it is evolved slowly as evidenced by
a very slight loss in weight, and also by the coloration or precipitation
of the lead sulphide from the lead syrup in the course of each of the
above experiments, a result which would not have obtained, were the

378 VELEY : SOME SULPHUR COMPOUNDS OF BARIUM.

hydrosulphide accidentally oxidised to the thiosulphate or tetra-

thionate.

Properties of a Solution of the Hydrosulphide.

A saturated solution of barium hydrosulphide when warm dissolves
sulphar with effervescence of hydrogen sulphide, and formation of a
dark red liquid. On further addition of sulphur, yellowish-red
crystals appear; on pouring the liquid from these, which analysis
proved to be a mixture, and allowing the liquid to covul slowly, red
crystals separate.

These on analysis gave results corresponding to a formula,
4BaS,,7H,0, probably BaS,,2H,0 slightly dehydrated, or mixed with
a small quantity of the anhydrous tetrasulphide.

Calculated
Analysis of crystals. for 4BaS,,7H,0.
Barium 46°67 per cent. 46°12
Sulphur 43°27
10°61

100-00

Atomic ratio of barium to sulphur = 1; 3°96.

The red solution of barium tetrasulphide is not decomposed by
hydrogen sulphide, either at ordinary temperatures or at that of
boiling water, differing in this respect from calcium pentasulphide.

Conditions of Formation of Barium Thiocarbonate.

In my former paper on the sulphur compounds of calcium, it was
shown that carbon disulphide was absorbed by calcium hydroxyhydro-
sulphide or the damp sulphide, but not by the anhydrous sulphide or
by the hydrosulphide. These conclusions were confirmed by similar
experiments with the corresponding bariam compounds. ‘Tests were
made in which hydrogen, containing a known proportion of carbon
disulphide was passed through (i) the solid hydroxide, (ii) the
anhydrous sulphide, (iii) a solution of the hydrosulphide, (iv) the
crystalline hydroxyhydrosulphide mixed with a large excess of water,
(v) the crystalline hydroxyhydrosulphide, and (vi) the mixture of coal
and barium sulphide (obtained on roasting the sulphate), slightly
damped with water. In experiments (i), (ii), (iii), and (iv), no carbon
disulphide was absorbed, and in (v) and (vi) the hydrogen was
completely purified from this substance. The negative result obtained
in (iv) is explained by the decomposition of the hydroxyhydrosulphide
into the hydroxide and the hydrosulphide (the presence of the latter

VELEY : SOME SULPHUR COMPOUNDS OF BARIUM. 379

in the resultant solution being detected by quantitative tests), both of
which, taken separately, were shown to be inactive.

The thiocarbonate, as described by Berzelius, is a yellow crystalline
solid, adhering firmly to the sides of the containing vessel; it is
sparingly soluble in cold, readily soluble in hot water, but is at the
same time decomposed. Several attempts were made to obtain this
substance in a state fit for analysis, but its preparation presents even
greater difficulties than that of the corresponding calcium compound.
Alcohol, dilute and absolute, ether and carbon disulphide, were tried as
solvents, but without success. Experiments were also made to
regenerate barium sulphide from the mixture of thiocarbonate and
hydroxyhydrosulphide, obtained by saturating the latter with carbon
disulphide vapour ; this material was heated to a high temperature in
a retort, the air of which was previously replaced by hydrogen. A
yellow mass was obtained, which was damped and again tested by
passing through it hydrogen containing carbon disulphide; but it was
found to be no longer active.

Summary.

The main points of the above paper are as follows :—
(i.) By the action of hydrogen sulphide on the hydrate of barium

hydroxide, there is formed barium sulphide and water, in accordance
with the equation Ba(OH).,H,O + H,S = BaS + 3H,0.

(ii.) By the action of hydrogen sulphide on barium oxide in
aqueous solution, there is formed a solution of barium hydrosulphide,
from which crystals of the composition Ba(SH),.,4H,O separate ; this
substance is far more stable than the corresponding calcium com-

pound.
(iii.) A warm saturated solution of barium hydrosulphide dissolves

sulphur to give a red liquid, from which crystals of barium tetra-
sulphide, prubably BaS,,2H,O, separate.

(iv.) Barium hydroxyhydrosulphide absorbs carbon disulphide with
formation of an unstable barium thiocarbonate.

XXXVIII.—Amidodiphenylsulphonic Acid and Azo-dyes from
Diphenyl.

By Tuomas Carne.iry, D.Sc., and James ScuiesetmMan, University
y , ’ )
College, Dundee.

In preparing dye-stuffs from diphenyl, either of two methods might
be adopted :—

(1.) To combine one of the hydroxydiphenyls, or a sulphonic acid
thereof, with a diazo-compound, thus—

C,,H,(OH):N : N-X:SO,H or ©,,H,(SO;H)(OH)N : N-X, &e.

(2.) To sulphonate amidodiphenyl, diazotise the resulting amido-
diphenylsulphonic acid, and combine the diazo-sulphonic acid so
obtained with a phenol, amido-compound, &c., thus :—

C,,H(SO;H)-N : N-X-OH.

Attempts to prepare dyes by the former method did not succeed,
the yield of hydroxydiphenyl obtained being very small. The second
process, however, gave very satisfactory results.

The diphenyl employed melted at 70°5°, and was chemically pure,
having been crystallised several times from alcohol. lt was converted
into the nitro-compound by treatment with concentrated nitric acid in
glacial acetic acid solution. By this method 23 grains of pure para-
nitrodiphenyl, melting at 112—113°, were obtained from 60 grams of
diphenyl, a considerable quantity of the ortho-compound being formed
at the same time. By reduction with nascent hydrogen, the above
quantity of paranitro-compound gave 16 grams of pure paramido-
diphenyl, melting at 49°.

Paramidodiphenylsulphonic Acid, CyHs(NH.)*SO;H (probably
NH,’C,H,C,HySO;H = 1:4; 1:4).—No sulphonic acid of amido-
diphenyl appears to have been previously obtained. To prepare
the above sulphonic acid, 1 part of amidodiphenyl (m. p. 49°) was
heated with 4 parts of strong sulphuric acid at a temperature of 130°
for half an hour. After cooling, the product was poured into cold
water, by which the sulphonic acid, being very insoluble, was preci-
pitated; it was then freed from any more soluble modification by
thoroughly washing with water. Paramidodiphenylsulphonic acid
melts above 300° with blackening and considerable decomposition. It
is practically insoluble in cold water, and scarcely soluble at all, even
on boiling.

Sodium Paramidodiphenylsulphonate, C,.H,(NH,)‘SO;Na + 2H,0.—
This salt was obtained by boiling the acid with pure sodium carbonate.

AMIDODIPHENYLSULPHONIC ACID, ETC., FROM DIPHENYL. 381

After filtering and allowing to cool, the sodium salt crystallised out
in colourless needles, about 1 cm. in length, and in general form
closely resembling crystals of potassium nitrate, but with a silky lustre
when crushed. For a sodium salt, it is but sparingly soluble in water.
The crystals, after thoroughly drying between filter-paper, contain
2 mols. H,O, both of which they lose at 100°. Analysis gave the

following results :—
Calculated,
(C\sH,pNSO,Na+2H,0). Found,
7°57
10°54
12°98

Barium Paramidodiphenylsulphonate, [C,,Hs(NH.)*SO;].Ba + 4H,O.

-This was prepared from the sodium salt by dissolving the latter in
water and adding barium chloride, when the barium salt separated in
the form of small needle-shaped crystals. These, after thoroughly
drying between filter-paper, contain 4 mols. H,0, all of which they
lose at 100°. The barium salt is only very sparingly soluble in water.
Analysis gave the following results :—

Calculated,
C24 HopN 8.0, Ba + 4H,0. Found.

18°53
8°66
11-40
We were unable to make a more satisfactory analysis of the barium
salt from want of material. The sample analysed possibly still
contained a small quantity of moisture.

The sulphonic acid described above appears to be the only one
formed by the action of sulphuric acid on paramidodiphenyl under
the conditions we have stated, for we were unable to detect the
presence of a more soluble acid in the aqueous extract.

The dyes referred to below were made from the above paramido-
diphenylsulphonic acid by the following general method :—

The sulphonic acid from 2 grams of amidodipheny], after admixture
with 2 c.c. strong hydrochloric acid and 20 c.c. water, was diazotised
by the addition, with frequent stirring, of 1 gram of potassium
nitrite dissolved in 3 e.c. of water. To this was added gradually, with
continual stirring, a cold alkaline solution of the requisite phenol.
The alkaline phenol solution was previously prepared by dissolving
06 gram of ordinary phenol, quinol, or resorcinol, or 1 gram of a@- or
B-naphthol in 5 c.c. water containing 1 gram caustic soda, and then
sufficient hydrochloric acid to make the liquid nearly neutral ; it was
subsequently made strongly alkaline by the addition of 3} grams of
sodium carbonate dissolved in a minimum of water.

CARNELLEY AND SCHLESELMAN:

The dye, with the exception of that from quinol, which was very
soluble, was thus precipitated, and subsequently collected, and washed
with cold water.

Sodium phenolparadiazodiphenylsulphonate,

HO-C,HyN : N-C,,H,SO;Na.
—This dye-stuff was obtained as above in the form of an amber-
coloured precipitate, which is only sparingly soluble in cold water,
but readily soluble on warming.

It dyes wool, and especially silk, a brilliant and very beautiful
yellow. To raw jute, it imparts an amber shade, but appears to have
little or no affinity for bleached jute yarn.

The barium salt was prepared by precipitating a very slightly acid
solution of the sodium salt with barium chloride. It forms a yellowish-
brown precipitate, which is almost insoluble in cold water. On analysis

Calculated,
(C,H ,3;N.8O04).Ba. Found.

15°79

7°36

Sodium resorcinolparadiazodiphenylsulphonate,
(HO).C,HyN ; N-C,,H,SO,Na,
is a deep brick-red crystalline precipitate, which becomes yellowish-
brown on drying. It is only sparingly soluble in cold, but readily
soluble in hot water. It dyes silk, wool, and raw jute a beautiful
orange, and bleached jute yarn a pink shade, the last fibre, however,
appears to have but little affinity for the dye.
The barium salt, obtained like that of the preceding compound, is a
dark-brown insoluble precipitate. On analysis it gave—
Calculated,
(C\g,H,3N.SO;).Ba. Found.
15°78

7°35

Sodium Quinolparadiazodiphenylsulphonate,
(HO).C.HyN : N-C,,HySO;Na.
—This compound is very soluble even in cold water, and was not
precipitated therefrom by the addition of salt, so that the dye was
not obtained in the solid state. Its solution in water is amber-
coloured, and dyes silk and raw jute a very pretty golden-amber. On
wool, the shade is very much duller, whilst bleached jute yarn does

not take up the dye at all.
The barium salt is also very soluble in water.

AMIDODIPHENYLSULPHONIC ACID, ETC., FROM DIPHENYL. 383

Sodium a-Naphtholparadiazodiphenylsulphonate,
HO-C,H,N : N-C,.H,SO,Na,
is a deep reddish-brown almost black precipitate, which is fairly
soluble in water even when cold. It dyes silk, wool, and raw jute a
deep reddish-brown. It has scarcely any affinity, however, for
bleached jute.

The barium salt is a dark-brown precipitate, which becomes black
on drying. It is somewhat soluble in water, much more so in fact
than that of the corresponding compound from f-naphthol. On
analysis, it gave—

Calculated.
(C..H,;0,SN2)2Ba. Found.
14°49
6°77
Sodium B-naphtholparadiazodiphenylsulphonate,
HO-CyHeN : N-C,.H.SO;Na.
is a bright red precipitate of the colour of red sealing-wax. It is
but sparingly soluble in cold, but readily soluble in hot water. It
dyes silk, wool, and raw jute a very brilliant red, and bleached jute
yarn a reddish-pink shade. It also dyes cotton, mordanted with tin
salt or with tannic acid, a bright red; the colour, however, is not fast
upon cotton, but is in great part removed by boiling water.

The barium salt is a dull red and almost insoluble precipitate,
which on analysis gave—

Calculated.
(C..H,;0,;SN.2)Ba. Found.
13°96
6°52

When the dye-stuff last described is produced in a liquor, which is
only made strongly alkaline with sodium carbonate after the mixture
of the naphthol with the diazodiphenylsulphonic acid, it is obtained
of a much darker and less brilliant red, and is very much more soluble
in water. In this form, it dyes silk a dark-red, and wool and raw jute
a brownish-red shade. An analysis of the barium salt gave results
corresponding with the formula (HO-C,H¢N : N-C,,Hs-SO;)2Ba.

All the above dyes are fast colours on wool and silk, and are not
affected either by boiling water alone, or by boiling water containing
half a per cent. of soap, except the yellow dye from phenol, which
becomes slightly paler after boiling with soap. In the case of the
reddish-brown dye from a-naphthol, the colour, though not weakened
by treatment with a boiling soap solution, becomes a purer brown, the
red shade being removed.

XXXIX.—The Combustion of Cyanogen.

By Harotp Dixon, M.A., The Duke of Bedford’s Lecturer in
Chemistry, Balliol College, Oxford.

A FEW years ago, on discovering that a dried mixture of carbonic
oxide and oxygen would not explode on the passage of an electric
spark through it, 1 made a few experiments on other dried gases, to
ascertain whether they showed the same peculiarity. While I found
that hydrogen and other gases containing hydrogen, even after long
contact with anhydrous phosphoric acid, always exploded with oxygen
under the influence of the spark, experiments made with dry
cyanogen gave results which inclined me to the belief that cyanogen,
like carbonic oxide, did not explode with oxygen when pure, but
was oxidised indirectly through the interaction of the small quantity
of aqueous vapour usually mixed with it. My experiments were
of two kinds. In the first place, a well-dried mixture of cyano-
gen and air was introduced into the eudiometer of a gas analyser.
The pressure being reduced to 100 mm., a spark from a Rubhmkorff
coil was passed through the mixture without causing an explosion.
The pressure was then raised 5 mm. at a time, and a spark was passed
after each increase of pressure. In this way, the pressure was
increased to 800 mm. without the spark causing an explosion in the
mixture. ‘The tube was then moistened, and the experiment repeated.
The spark caused no explosion at pressures Jess than 300 mm.; at
305 mm. pressure, the gases exploded with violence. In this experi-
ment, the cyanogen seemed to behave in a manner similar to that of
carbonic oxide, which when nearly dry is not exploded by a spark at
low pressures, but explodes when the pressure is raised. One striking
difference, however, was observed between the two cases. When the
pressure was gradually increased until the spark caused the explosion
of the carbonic oxide and oxygen, the inflammation was _ noise-
less and extremely slow, the blue disc of flame taking some seconds
to travel down the eudiometer ; but when the cyanogen and oxygen
were exploded under 305 mm. pressure, the inflammation appeared
instantaneous, and was accompanied by the bright flash, and the
“‘ ping” on the mercury characteristic of violent explosions in eudio-
meters. In the second place, two mixtures of cyanogen and oxygen
were kept standing over mercury in contact with anhydrous phos-
phoric acid for several days in bent eudiometers. On testing these
mixtures with sparks from a Holtz machine, I found that neither of
them exploded. Tested with the discharge from a coil, one mixture
exploded violently, the other was unaffected. I regarded these pre-

DIXON: THE COMBUSTION OF CYANOGEN. 385

liminary experiments as affording primd facie evidence of the non-
inflammability of pure cyanogen in oxygen. Other work kept me
from examining the reaction more in detail at the time.

On returning to the subject last autumn, I discovered that the
explosion of cyanogen and oxygen is independent of the presence or
absence of steam. The explosion depends solely on the nature of the
spark. A strong spark causes the mixture to explode violently,
whether wet or dry; a weak spark may be passed through the mixture
wet or dry, without apparent effect. Although these experiments form
part of a more general investigation (not yet completed) of the
phenomena of gaseous explosions, Iam glad to take an opportunity
of correcting before the Chemical Society, the error to which I
formerly leaned.

The cyanogen employed in the following experiments was prepared
from mercuric cyanide, powdered and dried at 120°. The dry powder
was placed in a piece of combustion tubing, closed at one end, and
drawn out into a delivery tube at the other. The gas was collected
over dry mercury, and kept in contact with anhydrous phos-
phoric acid. The oxygen was prepared in a similar manner from
powdered potassic chlorate. Mixtures of the dry cyanogen and
oxygen were made in short straight eudiometers, furnished with
platinum wires fixed at different distances apart. In one tube
the wires approached to within about 0°25 mm., in a second
to within 0°5 mm.; in the others the distance between the wires
varied from 1 mm. to 3 mm. These eudiometers were filled while
hot with hot mercury; after the gases had been introduced, short
sticks of phosphoric acid were passed up through the mercury into
the mixtures. The tubes were then allowed to stand for a fortnight
before testing. On passing sparks from a Holtz machine through the
mixtures none of them exploded. The smallest spark (0°25 mm.) from
the Holtz machine exploded mixtures of hydrogen and oxygen and
carbonic oxide and oxygen (damp) under the same conditions. On
passing the discharge from a Ruhmkorff coil, no explosion occurred
in those tubes in which the platinum wires were close together—from
0-25 to 1 mm. apart; the mixtures in the other tubes exploded
violently. Comparative experiments were then made with the gases
moist. Sparks from the Holtz machine were passed through all the
tubes without explosion; the Ruhmkorff discharge failed to explode
the mixture in the three tubes in which the platinum wires were
close together, but it caused a violent explosion in the other tubes.*

* Two of these tubes filled with the same mixture were exhibited before the
Society on March 18th. Sparks from a Ruhmkorff coil were passed between wires
1 mm. apart in one tube without causing explosion. In the second tube the wires
were 3 mm. apart. The first spark passed between them caused a violent explosion.

VOL. XLIX. 2D

386 DIXON: THE COMBUSTION OF CYANOGEN.

Since the conditions of temperature and pressure were, as nearly as
possible, similar in the two series of experiments, it seemed probable
that the explosion or non-explosion of cyanogen and oxygen was
determined by the nature of the spark, and not by the presence or
absence of moisture.

To make sure that a thoroughly dried mixture of cyanogen and
oxygen would explode with a powerful spark, I kept the gases in
contact with a considerable quantity of anhydrous phosphoric acid for
eight weeks. The tube was occasionally shaken, to break up the
loosely packed sticks of oxide, and so offer fresh drying surfaces to
the gas. On passing a large spark through the gases after the two
months’ drying, the mixture exploded with the usual violence. In
another experiment, the cyanogen was kept ten days in contact with a
stick of potash, and was then mixed with dry oxygen in a eudiometer.
After the mixture had stood several days in contact with anhydrous
phosphoric acid, it exploded violently on the passage of a strong
spark. This experiment showed that the explosion of cyanogen was
not dependent on the presence of a small quantity of hydrocyanic
acid or other substance absorbable by potash.

But it may still be objected that the drying by phosphoric acid,
however prolonged, is not really complete, and that the trace of aqueous
vapour remaining is sufficient to bring about the explosion, once the
reaction is started by a sufficiently strong spark. Supposing, it may
be urged, we possessed in chemistry no more powerful drying agent
than potash, we should find that a mixture of carbonic oxide and
oxygen, dried as thoroughly as our means permitted, would withstand
without explosion a spark from a Holtz machine, but would explode
when submitted to the discharge of a Ruhmkorff coil. What happens
in the case of carbonic oxide dried by potash may happen also in the
case of cyanogen dried by phosphoric acid, and no positive evidence
has been adduced to show that pure cyanogen and oxygen are capable
of exploding, but only that these gases still explode in the presence
of an exceedingly minute trace of water. The argument would be
ananswerable if it had not been shown that the rate of combination
of carbonic oxide and oxygen increases rapidly with the addition of
aqueous vapour, whereas the rate of combination of cyanogen and
oxygen is not increased, but slightly diminished, by the addition of
aqueous vapour to the very dry gases. Exact measurements of the
initial velocity of explosion of dry and wet mixtures of cyanogen and
oxygen leave little room for doubt that the addition of aqueous
vapour only impedes the reaction between the two gases.

The explosion-rate of cyanogen and oxygen was determined by
means of the pendulum chronograph employed in my previous experi-
ments on carbonic oxide and oxygen. The explosion tube was similar

DIXON: THE COMBUSTION OF CYANOGEN, 387

in general design to the one used in the earlier research, but the taps
and joints of the new tube were made of hard steel instead of gun-
metal, which was found worn away where it had been continually
subjected to the molecular bombardment of the heated gases. The
explosion-tube AA was 10 feetlong. Near one end of it, B, two insu-
lated platinum wires were let into the tube, so that a distance of about

Arrangement of Explosion Tute.

388 DIXON: THE COMBUSTION OF CYANOGEN.

3 mm. separated their points. These wires were connected up with
the secondary coil of a large Ruhmkorff. Near the other extremity
of the tube, C, were fitted two insulated silver pieces, one on each side
of the tube, each connected with a platinum wire passing through a
vulcanite block to the outside of the tube. Before each experiment, a
narrow strip of silver foil was soldered on to the insulated silver
pieces, forming a slender “ bridge” across the tube. This silver bridge
carried a current from one Grove cell to an electro-magnet, which, so
long as the current passed, held down a style against the pull of a
light spring; when the explosion, passing down the tube, smashed
the silver bridge, the electro-magnet was unmade, and the style,
released, sprang upwards at the moment the circuit was broken. A
heavy pendulum, carrying a smoked glass plate, passed in its swing
in front of the style, which left its trace on the blackened surface.
The moment at which the circuit was broken was thus recorded on the
plate. At the bottom point of its swing the pendulum broke another
circuit connected with the primary coil of the Ruhmkorff. At the
moment this primary circuit was broken the induced current in the
secondary coil caused a spark to pass between the platinum wires in
the explosion-tube, firing the mixtures of gases. The exact moment
of firing was thus known. While the explosion was passing along the
tube between the firing point and the silver bridge, the style con-
tinued to trace an are of a circle on the moving plate, until, on the
explosion reaching the bridge, the circuit was broken, and the style
was released from the electro-magnet. The distance travelled by the
pendulum between the passage of the spark and the rupture of the
bridge—translated into time by means of a standard tuning fork—
gave the mean velocity of explosion of the gaseous mixture. One
considerable error had to be corrected, the retardation of the electro-
magnet, which varied with the strength of the current. This was
eliminated by means of a preliminary experiment, made immediately
before the explosion. The electro-magnet (detached from the silver
bridge) was connected up so that its circuit was broken by the pendu-

Plan of the cap C. Section of the cap C.

DIXON : THE COMBUSTION OF CYANOGEN. 389

lum instead of the primary circuit of the Ruhmkorff. This moment
was recorded by the style with the error due to retardation. Without
altering the position of the electro-magnet and style, the wires were
readjusted and the explosion made. The second mark made by the
style gave the moment at which the silver bridge was broken, with
the error due to retardation; and since the previous mark gave the
moment at which the mixture was fired, also with the error due to
retardation, the distance between the two marks gave the duration
of the explosion independently of the error, which affected equally
the position of both marks.

The gases were collected and measured in a dry iron cylinder over
mercury in an annular iron tank. The cyanogen was passed directly
from a glass retort through a U-tube filled with potash into the holder,
previously charged with dry oxygen. The volume of the mixture
amounted to half a cubic foot, so that all the experiments could be
made with the one mixture. The gases, after mixing, were slowly
forced into the dry explosion-tube through two vessels packed with
asbestos coated with anhydrous phosphoric acid. Two determinations
of the rate of explosion were made with the gases dried in this
manner. Ina third experiment, the gases were passed into the dry
explosion-tube through a U-tube packed with fragments of potash ;
in two other experiments, the gases were made to bubble through
water at 15° before entering the tube. The temperature of the room
was 15°, and the barometer stood at 763mm. In the following table,
the result of these experiments is given, and for the purpose of com-
parison a table of the similar experiments with carbonic oxide and
oxygen is placed beside it. It must be borne in mind that the rates
here given are not the constant rates of the explosive waves of car-
bonic oxide and of cyanogen. The experiments with carbonic oxide
were made in a tube about 3 feet long, of which one-third only was
traversed at the full speed of the “explosive wave.” In the cyanogen
experiments, probably only the first foot was traversed with a con-
stantly increasing velocity, and the remaining 9 feet at full speed.

Mean Initial Rate of Explosion.

1. Carbonic oxide and oxygen. 2. Cyanogen and Oxygen.

| Rate. Meters : | Rate. Meters
per sec. Hygrometric state. per see.

Hygrometric state.

Dried by P,O;.......+. 36 Dried by PO5.. 0... 0.05 813
ia ie 119 v4 da 811

” 2 ”
Saturated at 10°C. ...... 175 <> Mn as4u cues 808
i Ta webees 244 Saturated at 15° C. ..... 752
- TP me sacees 317 “ 741

390 DIXON: THE COMBUSTION OF CYANOGEN.

I would call special attention to the rapid increase in the explosion-
rate of carbonic oxide and oxygen as a trace of water is added to the
nearly dry mixture, compared with the almost inappreciable fall in
velocity as a trace of water is added to the nearly dry mixture of
cyanogen and oxygen. Unless we are prepared to admit that the
immeasurably small quantity of steam remaining in the gas after
passing through the phosphoric acid is enough, or more than enough,
to effect the combination of the cyanogen and oxygen at their mawimum
rate, these results seem conclusive that the explosion of cyanogen and
oxygen is not brought about by the interaction of a trace of steam,
but by a direct auction between the gases under the conditions of the
experiment.

In a paper (this Journal, Trans., 1885, p. 575) read before the Society
last year, I described the action of a heated coil of platinum wire
on a dry mixture of carbonic oxide and oxygen. When the coil was
raised to redness in the thoroughly dried mixture, the wire glowed
intensely for a minute or more, and after cooling the whole of the
carbonic oxide was found to be oxidised. No trace of flame was
visible in the tube during the combination. A similar experiment
made with cyanogen and oxygen gave a curiously analogous result.
The tubes employed were the same as those described in my former
paper.

When the coil of platinum wire was raised to dull redness by the
current, and the circuit was broken, the coil quickly cooled. But
when the coil was raised to full redness, it glowed brightly, and
continued glowing for half a minute after the circuit was broken. No
trace of flame could be detected round the wires. On cooling, the
tube was found filled with orange vapour.

The experiment was repeated with the same result, both when the
cyanogen and oxygen had been thoroughly dried by long contact with
phosphoric acid and when the gases were moist. In the latter case,
the orange fumes disappeared in about an hour, and a white crystal-
line deposit appeared on the mercury. The formation of orange
vapour was observed even when the oxygen was taken in considerable
defect, but the amount formed was less in this case than when the
oxygen was in the normal proportion. On analysing the residual gases
left after the coil had been heated in a dry mixture of cyanogen and
oxygen, it was found that the combination was not quite complete.
Of the cyanogen acted on, about three-fourths had been converted
into carbonic acid and one-fourth into carbonic oxide.

This slow combustion of cyanogen in presence of heated platinum
was first observed by Davy. In the Phil. Trans. of 1817 (p. 80), he
says :—

“The chemical changes in general produced by slow combustion

WITT : THE EURHODINES. 391

appear worthy of investigation. A wire of platinum introduced
under the usual circumstances [heated to low redness] into a mixture
of prussic gas and oxygene in excess became ignited to whiteness, and
the yellow vapours of nitrous acid were observed in the mixture.”

My observations entirely confirm those of Davy. No explosion is
produced by the heated wire even when the gases are mixed in the
theoretical proportions, and the reaction is independent of the dryness
or dampness of the mixture.

The non-explosion of cyanogen and oxygen under the influence of
a small spark would appear at first sight to be due to an insufficient
temperature being reached. Berthelot (Compt. rend., 95, 955) has
shown that when cyanogen is passed through a tube heated to bright
redness, it slowly undergoes decomposition, whereas a strong stream
of sparks rapidly splits it up into carbon and nitrogen; and again
V. Meyer and Goldschmidt (Ber., 1882, 1161) have shown that while
the density of cyanogen remains normal between 100° and 800°,
decomposition occurs at 1200°. It would seem therefore plausible
to suppose that the compound cyanogen is not acted on by the
oxygen, but that when broken up by the spark the liberated carbon
is oxidised, and the heat so generated breaks up more cyanogen,
and soon. It is, however, hard to reconcile with this view the fact
that white-hot platinum, though causing a rapid combination of
cyanogen and oxygen, yet produces no explosion in the mixture; and
again, when a stream of small sparks is passed for several hours
through a mixture of cyanogen and oxygen, partial combination
occurs, and the mixture of gases becomes yellow. The temperature
of the small spark is therefore sufficient to produce some action
between the cyanogen and the oxygen. I hope to make further
experiments with a view to decide this question.

XL.—The Eurhodines, a New Class of Colouring Matters.
By Orro N. Wirr, Ph.D.

One of the methods which lead to the synthesis of dyestuffs consists in
the jointoxidation or rather dehydrogenisation of mixtures of various
amines, or of amines with phenols; several classes of new colour-
ing matters may thus be obtained according to the amines employed.
Thus, the oxidation of a paradiamine with one molecule of a
monamine leads to the formation of a member of the indamine-group,
whilst one molecule of paradiamine with two molecules of a mon-

392 WITT: THE EURHODINES,

amine yield a safranine, if subjected to the action of powerful
oxidisers. Again, a mixture of a paradiamine and a metadiamine in
molecular proportions yields on oxidation a compound belonging to
the toluylene-blue-group, and if still more hydrogen be removed, a
condensation into a derivative of the toluylene-red group is the result.
Mixtures of paradiamines and phenols, if oxidised in alkaline or
neutral solution, give rise to dyestuffs of the indophenol class, whilst
more powerful oxidation in acid solutions results in the formation of
substances similar to those discovered and described by R. Meldola.
All the classes of dyestuffs enumerated have been studied more or
less completely in some of their most important representatives, and
we even know the exact chemical constitution of some of them.

Now, in all these reactions, one of the ingredients is always a
paradiamine. No attempt has been made to replace this by other
substances capable of similar reactions. The reason for this is
probably the fact that most of these reactions have not been dis-
covered in the form here given to them. As a rule, certain colour-
reactions were obtained with amido-azo- or nitroso-compounds, and
it was only recognised on subsequent examination, that the diamine
formed by the reduction of these azo- or nitroso-derivatives must be
considered as the first cause of the colour-reaction itself. Now as all
amido-azo-compounds and nitrosamines known up to a few years ago
belonged to the para-series of benzene-derivatives, the constant
recurrence of paradiamines in these reactions was very natural.

In 1882, however, the first orthamido-azo-compound was dis-
covered by Nélting and Witt (Ber., 17, 77), who obtained it by
heating pure diazoparamidotoluene with paratoluidine hydrochloride
in paratoluidine solution on the water-bath. This substance
crystallises in needles (from acetone) or leaflets (from benzene) of
a brillant orange colour, and melts at 118°5°; on reduction, it splits
up into paratoluidine and orthotoluylenediamine of the m. p. 88°5°,
the constitution of which is known to be—

The new amido-azo-compound is therefore, of necessity,

4 \Nx:xd \
CHK DN:INK DCB.
NH,

: If this substance is treated with hydrochloric acid in alcoholic solu-
tion, it yields a hydrochloride which, unlike the salts of paramido-azo-

A NEW CLASS OF COLOURING MATTERS. 393

derivatives dissolves with a green coloration. In the solid state, it
forms needles of pale straw colour. If this salt is acted on by primary
mo namines, such for instance as aniline, it produces, like the salt of a
paramido-azo-compound, a dye of the induline-group; now it has
been shown by Thomas and Witt (Trans., 1883, 112), that in the
induline reaction the amido-azo-compound employed plays no other
part than that of dehydrogenising the primary monamine acted upon,
and if new proof were needed for the correctness of that view, it would
be found in the very fact that an orthamido-azo-compound, acting
on aniline, produces the same final result as that obtained with
paramido-azo-compounds. But on the other hand, this experiment
shows that no reply can be obtained in so simple a manner to the
question whether the orthodiamine (formed by the reduction of the
orthamido-azo-compound) is capable of forming dyes like its para-
isomeride. An unexpected answer to that question was, however,
afforded when a-naphthylamine was substituted for aniline in the
above experiment; a new dyestuff was obtained, which was so per-
fectly unlike any colouring matter hitherto observed, that it could at
once be recognised as belonging to a new class of compounds, the
parent-substance of which could only be the orthotoluylenediamine.
This hypothesis was confirmed when other orthamido-azo-compounds
were treated with a-naphthylamine hydrochloride. In all cases
colouring matters were formed closely resembling the one first
obtained. Itthen became interesting to know how chrysoidine would
behave under similar circumstances. This well-known dyestuff is, as
may be seen from its constitutional formula—

4 N\xn-wt NS
< Dw >NEa,
NH,

both a para- and an ortho-amido-azo-compound. If its hydrochloride
is heated with «-naphthylamine at 130°, a dyestuff belonging to the
new class, and showing all its peculiarities, is obtained; it can,
therefore, no longer be doubted that the ortho-position of the amido-
and the azo-group is the sole cause of the new reaction, which was
carefully investigated. The results obtained I have now the pleasure
of laying before the Society.

Action of a-Naphthylamine on Orthamido-azotoluene.

If 22°5 grams (1 mol.) of orthamido-azotoluene (m. p. 118°5°)
is mixed with 17°9 grams (1 mol.) of 2-naphthylamine hydrochloride,
and 50 grams of phenol added, a thick solution of dark emerald-
green colour is obtained. The phenol acts merely as a convenient

394 WITT: THE EURHODINES,

solvent, and may be replaced by a-naphthylamine, which is, however,
not so easily got rid of afterwards. The green coloration is the
result of the formation of the hydrochloride of the azo-compound, the
salt of the a-naphthylamine yielding up part of its acid to the azo-
compound. If this mixture be placed on the water-bath, the green
coloration will gradually disappear and be replaced by an intense and
brillant scarlet colour. The reaction is then finished and the new
dyestuff may be isolated from the mixture. Two different methods
may serve equally well this purpose. The first consists in adding to
the warm mixture a sufficient quantity of toluene to precipitate all
that is insoluble in that solvent. A bright scarlet precipitate,
consisting of the hydrochloride of the new dyestuff, is the result.
This is collected on a filter, washed with ether, and then igtro-
duced into boiling water, acidulated with hydrochloric acid, and
dissolved by prolonged boiling. The filtered liquid on cooling
deposits long interlaced needles of the pure hydrochloride, which may
be collected, washed with very dilute hydrochloric acid, and dried.

Another method of working up the crude melt consists in adding
sufficient caustic soda solution to dissolve all the phenol and collecting
the yellow precipitate which separates. By treating it with dilute
hydrochloric acid, it is transformed into hydrochloride, which may
then be treated as before.

If the pure hydrochloride obtained by either of these methods is
decomposed by boiling weak ammonia solution, the free base separates
as an amorphous precipitate of brilliant yellow colour, which becomes
distinctly crystalline on prolonged boiling. It is collected, dried,
and recrystallised either from aniline or from phenol, with addition of
alcohol. By this means, it is obtained in dark orange glistening
needles, which on analysis were found to have the composition

C,;Hi:N2.
Found.

Calculated
for C,;H,;N3.

To this substance I have given the name of eurhodine, from cd
and pwéos, to recall the red colour of its salts and the eminent crys-
tallising power of the free base.

In order to establish with greater certainty the molecular weight of
eurhodine, its hydrochloride was subjected to analysis. This salt was
prepared in a state of absolute purity, and in the shape of long, soft,
brilliant garnet-coloured needles, by adding dilute hydrochloric acid
to a solution of the pure base in acetic acid, filtering, washing with

wee el lll lle

77)

A NEW CLASS OF COLOURING MATTERS. 395

dilute hydrochloric acid, and drying in a vacuum over quicklime and
sulphuric acid. It contains 1 mol. H,O, which cannot be driven out
by heating without partial decomposition of the salt. The elementary
constituents of this salt were determined as well as the quantity of
free eurhodine base it contains.*

Found.

Calculated for
©),H,3N3,HCl + H,O.

11°33
C,,H,N; 82°60 82°77

The sulphate, oxalate, and tartrate of eurhodine closely resemble the
hydrochloride. The nitrate is absolutely insoluble even in boiling
water. The acetate could not be obtained in a dry state, but its
existence is proved by the fact that eurhodine dissolves with the
scarlet colour of its salts in acetic acid.

Eurhodine and its salts show several most striking reactions. In
concentrated sulphuric acid they dissolve with intense red colour.
This solution on being carefully diluted with water becomes of an
intense green. Further addition of water produces a change to
scarlet, and the crystalline normal sulphate separates from the
solution. Crystallised eurhodine is almost insoluble both in alcohol
and in ether; but if it is precipitated in its amorphous condition by
adding ammonia to one of its salts in the cold, and then shaken with
ether, it is freely taken up by this solvent. The solution is of a bright
yellow colour, and shows an intense green fluorescence, quite as strong
as that of fluorescein but of a more yellow shade. This solution
shows a spectrum similar to that of sodium fluorescein. Free eurho-
dine, as well as all its derivatives, show a remarkable indifference to
high temperatures. Unlike all other dyestuffs—which are decomposed
by dry distillation—eurhodine is volatilised without decomposition.
Its vapour on cooling condenses to a voluminous cloud of micro-
scopic crystals. This sublimate strongly resembles the one obtained
from certain anthraquinone-derivatives. A few milligrams of
eurhodine may be made to fill a large space in this sublimed con-
dition.

All the salts of eurhodine are partly decomposed by pure water.
In the presence of an excess of acid, however, they are quite stable.

Eurhodine is a true dye. On silk it dyes a scarlet shade in an acid

* In my preliminary notice (Ber., 19, 443) the calculated data were erroneously

given as found, and vice versd. This error was subsequently corrected (Ber., 19,
592).

396 WITT: THE EURHODINES,

bath. Its tartrate mixed with tannin and acetic acid, suitably
thickened, and printed on cotton mordanted with Turkey-red oil,
steamed and washed, produces a red similar to Turkey-red. But the
shades produced both on silk and cotton are changed to yellow
on prolonged washing in ordinary calcareous water. The original
red may be restored by acids.

The formation of eurhodine may be easily accounted for by
assuming that equal molecules of orthotoluylenediamine and a-naph-
thylamine are condensed by the elimination of 6 atoms of hydrogen,
the latter being taken up—as is always the case in such reactions—
by the reduction of the azo-group—

C;HyN:, orthotoluylenediamine,
CyoH,N;, «-naphthylamine,

Cy HN;

C,,H,,N;, eurhodine.

If this hypothesis be correct, the other half of the amidoazo-
compound, paratoluidine, must be found free in the crude melt.
This is actually the case; paratoluidine was shown to be present in
abundance in the mother-liquors from the preparation of the new
dye.

Eurhodine is a substance of such marked properties, and its forma-
tion is so simple and sure, that it may be utilised as a test, indicating
the relative position of any new amidoazo-compound. It is only neces-
sary to heat a trace of the latter with a few crystals of a-naphthyl-
amine hydrochloride and a drop of liquefied phenol, to add an excess of
caustic soda and to shake up with ether. If a yellow solution with
bright-green fluorescence be obtained, one may be certain that the
amidoazo-compound employed belongs to the ortho-series.

Having established the mode of formation as well as the properties
of eurhodine, I made some experiments with a view to determining
the constitution of the new dye.

Action of Acetic Anhydride.

If acetic anhydride be poured on to eurhodine crystals contained in
a dry flask, solution takes place. But very soon crystals begin to
separate, and the whole mixture solidifies to a crystalline cake.
The acetyl-derivative thus formed may be collected on a filter and
washed with alcohol. It forms fine felted needles of straw-yellow
colour, which are split up into acetic acid and the free base by the
action of strong mineral acids.

A NEW CLASS OF COLOURING MATTERS. 397

The formation of an acetyl-derivative makes it probable that eurho-
dine contains a primary amido-group, and this probability became a
certainty by the examination of the

Action of Acids on Eurhodine.

If eurhodine or one of its salts is heated with dilute hydrochloric
or sulphuric acid in a sealed tube, it undergoes a complete and very
remarkable decomposition ; 3—4 hours at 180° suffice to make every
trace of eurhodine disappear. In its stead, yellow crystals separate
from the colourless supernatant liquid. On evaporating the latter
to dryness, no organic substance could be detected. The entire
residue consisted of ammonium chloride or sulphate. The ammonia
could only have been formed by an amido-group being eliminated from
eurhodine by the action of the acid.

The crystals formed in the tube consisted of a new substance, which
possesses both basic and acid properties. It is consequently the
phenol corresponding with the primary amine eurhodine. If this be
the case, its composition must be C,;H,,N.0. Analysis proved this
supposition to be correct.

Found.

Calculated for —
II.

78°2
49
— 10°5

To this phenol of the eurhodine-group, I have given the name of
eurhodol.

Eurhodol is very insoluble in all solvents with the exception of
aniline and phenol. In the amorphous state it dissolves pretty freely
in ether ; this solution is of a yellow colour, but it is not fluorescent.
If slowly crystallised from aniline or phenol, the crystals are of
an indistinct orange colour. If, however, the phenolic solution be
cooled rapidly, the crystals will be of a bright yellow shade, whilst the
aniline solution, by rapid cooling, yields crystals of a vermilion colour.
This phenomenon is due to dichroism of the crystals. They possess
both red and yellow planes, which are differently developed according
to the solvent used.

Eurhodol may be volatilised almost without decomposition. It forms
a yellow sublimate similar to that of eurhodine.

In concentrated sulphuric acid, eurhodol dissolves with a dark red

colour; water precipitates the sulphate from this solution in yellow
flakes.

398 WITT : THE EURHODINES,

Concentrated caustic soda solution dissolves eurhodol with an orange
colour, and on cooling, a crystallised sodium salt separates. From
the dilute soda solution, free eurhodol may be precipitated by the
addition of acetic acid.

It thus became evident that eurhodine is the amine, eurhodol the
phenol, of a substance C,;H,,N2, which I tried to discover by study-
ing the

Action of Nitrites on Eurhodine.

If sodic nitrite be added to an acid solution of eurhodine, reaction
takes place, and a diazo-compound is formed, as may be seen from
the action of the liquid on sodium #-naphthol. It is impossible,
however, owing to the great insolubility of all eurhodine-derivatives,
either to finish the reaction or to isolate a diazo-compound from the

mixture.

Better results were obtained by the action of alkylic nitrites on
alcoholic solutions of eurhodine salts. If the acetate is dissolved in
alcohol, ethyl nitrite added, and the mixture boiled for two hours on
the water-bath, red crystals separate. This substance, which in all
probability is the amidoazo-compound of the hypothetical C,,;H,.N:,

has not yet been fully examined. It is feebly basic, and dissolves
with dark colour in sulphuric acid; on dilution with water, the colour
changes to red, and a sulphate separates in crimson flakes.

In presence of a mineral acid, the reaction takes place somewhat
differently. Small quantities of the red substance are formed, but the
greater part of the eurhodine is acted on with elimination of nitro-
gen; the liquid soon becoming yellow. The alcohol is then distilled
off, and the residue treated with dilute (15 per cent.) hydrochloric
acid, which takes up a new base. On adding an alkali, the base is
precipitated, and may be collected and recrystallised from a mixture
of glacial acetic acid and alcohol. It is thus obtained in long,
glistening needles of bright lemon-yellow colour. On analysis these
needles proved, however, not to be the expected base C;;,H,,N.. The
data obtained showed that it contained oxygen, and that its composi-
tion was C,.H,,.N,O.

Found.
Calculated for -— sO

CyoHygN0. 1. I.
79°16 78°96 —
5°56 59200

9°72 — 10°23

5°56 — —_—

This new substance melts at 175°. It is volatilised almost without
decomposition at a high temperature. It is easily soluble, with yellow

A NEW CLASS OF COLOURING MATTERS. 399

colour and a faint green fluorescence, in benzene, ether, and acetic acid,
less so in alcohol.

Alcoholic potash, even at 150°, is without any action on the new
compound. It is easily soluble in concentrated sulphuric acid, with
red colour. It is dissolved by concentrated hydrochlore acid with an
orange coloration, and after a while, brown, bronzy crystals of the
hydrochloride separate from the solution. If this solution is heated
at 180° in sealed tubes, the compound is transformed into an acid sub-
stance resembling eurhodol.

It became evident by these reactions, as well as from the analytical
results, that the action of ethyl nitrite had not resulted in the substi-
tution of hydrogen for NH, but that, as is sometimes the case, the
ethylic ether of the corresponding phenol had been formed, and that
the rational formula of the new substance must be written

C,;H,,N,0°C.H;.

Although thus far my experiments with a view to isolate C,,H,.N2,
the parent-substance of the whole eurhodine-group, had been failures,
there was still something to be learned from them respecting the
nature of the compound ©,;H,.N2. It will be remembered that phenolic
ethers invariably show a strong resemblance to the corresponding
hydrocarbons. Thus anisoil is a compound strongly resembling ben-
zene. In all probability, therefore, the properties of C,;H,.N, would
be very similar to those of the substance C,;H,,N,0-C.H;.

Now, in going over what is known about such substances, we find
that a compound, C,;H,.N2, has already been prepared by Hinsberg
(Ber., 18, 1228), who obtained it by acting on orthotoluylenediamine
with 8-naphthaquinone, and to which he ascribes the constitutional
formula—

SF ee,

which 1 prefer to write, for reasons set forth hereafter—

400 WITT: THE EURHODINES,

This compound, naphthylenetoluquinoxaline, belongs to the class
known as quinoxalines, the first and typical representative of which
is phenylenequinoxaline—

ei \

oe }
N

also prepared by Hinsberg (Ber., 17, 320) by the action of glyoxal

on orthophenylenediamine.

I have prepared Hinsberg’s naphthylenetoluquinoxaline, and com-
pared its properties with those of my ether, obtained by the action of
ethylic nitrite on eurhodine. The two substances are remarkably
similar, and it became thus highly probable that eurhodine is amido-

naphthylenetoluquinoxaline.

I was anxious, however, to prove by some other method that the
introduction of the amido-group into the molecule of a quinoxaline
was sufficient to form a dyestuff of the eurbodine character. I have,
therefore, tried to obtain by synthetical means other amidoquinoxalines.
My endeavours in that direction were successful.

Phenanthro-eurhodine.
Glyoxal, as may be seen by a glance at its constitutional formula—
O:C-H
O. ba
is the prototype of a long series of compounds which may be described
as orthodiketones, and which may be derived from glyoxal by its

hydrogen atoms being displaced by other radicals. Thus, by sub-
stituting phenyl for hydrogen, we obtain benzil—

O ; C-C,H;
| ;
O i C-C,H;

whilst the introduction of the diphenylene-group gives rise to the
formation of diphenylene diketone, or phenanthraquinone—

0: C—C,Hy

a oe
O : C—C,Hy
Now all these substances, when exposed to the action of ortho-

diamines, act on the latter like glyoxal, producing substituted
quinoxalines. For my purpose, I had, therefore, to react with any or

A NEW CLASS OF COLOURING MATTERS. 401

either of these substances on an amidorthodiamine, or, in other words,
on a triamine, two amido-groups of which are in the ortho-position
relatively to each other. Such a substance is 1: 2: 4-triamido-
benzene—

a substance which is formed by the reduction of dinitraniline or
chrysoidine. I soon found that it was not necessary for my purposes
to isolate this highly unstable substance. It is sufficient to act with
any orthodiketone on the crude reduction product of chrysoidine in
order to obtain the corresponding eurhodine.

If 2 grams of chrysoidine crystals are dissolved in 50 grams of
glacial acetic acid, and an excess of zinc-dust added to this solution,
reduction takes place instantaneously, and a colourless liquid is
obtained. This is poured off from the zinc-dust (not filtered to avoid
the contact of the air), and mixed with the solution of 1 gram of
phenanthraquinone in 50 grams of glacial acetic acid. The mixture
assumes _ instantaneously an intense red colour, and the new
eurhodine is formed. By adding a large quantity of water, the new
compound is precipitated in yellowish-brown flakes, which become
bright yellow on prolonged boiling. The precipitate is then
collected, and recrystallised from phenol and alcohol. In this way
15 grams of small brown crystals are obtained, which on being
subjected to analysis proved to have the expected composition

Found.
Calculated —_—_—r
for Cop H,3N3. I. i.
81°4 —_—
4°7 —
= 145

This substance, the constitution of which is undoubtedly—
N

»
yn

iT C—C, H,

is shown by all its reactions to belong to the eurhodine cla:s of colour-
ing matters.
Its solution in concentrated sulphuric acid is red, and on being
VOL. XLIX. 25

402 WITT: THE EURHODINES,

diluted with water becomes yellowish-green, and then again red. Its
salts are highly insoluble, and of a carmine-red colour. On adding
an alkali to their solution, the free base is precipitated in its amorphous
state, and is taken up by ether with yellow colour and the charac-
teristic green fluorescence of the eurhodines. The free hase is very
insoluble in all solvents, except aniline and phenol. It is volatilised
at a high temperature almost without decomposition, and a woolly
sublimate is formed in the cooler portions of the tube.

Similar eurhodines were obtained by acting on 1:2: 4-triamido-
benzene with any of the orthodiketones capable of forming quin-
oxalines. The experiment was tried with gyloxal, dioxytartaric
(carboxytartronic) acid, benzil, 6-naphthaquinone, and isatin. A red
dye of the eurhodine character was invariably the result.

The reaction holds good even for polyketones. Thus leuconic acid,
which was proved by R. Nietzki to be pentaketopentamethylene—

O
C

F in
0:C C:O
| =|
0:C—-—C:0O

if treated with triamidobenzene, produces a violet colouring matter,
which becomes brown on addition of alkalis.

With the results thus obtained, it could no longer be doubted that
eurhodine is amidonaphthylenetoluquinoxaline, but there was still one
point to elucidate.

Only one naphthylenetoluquinoxaline has hitherto been mentioned,
and only one has been prepared by Hinsberg. But, according to our
present theoretical views on the constitution of naphthalene, a second
isomeride may be predicted. This becomes evident from the following
consideration.

Two orthonaphthaquinones are capable of existence, one of which
only has hitherto been prepared by Stenhouse and Groves (Chem.
Soe. J., 1877, ii, 47 ; 1878, 415).

”
O

8-Naphthaquinone of Missing
Stenhouse and Groves. isomeride.

Now the second of these, if accessible, would be equally capable of
producing a naphthylenetoluquinoxaline if exposed to the action of

A NEW CLASS OF COLOURING MATTERS. 403

orthotoluylenediamine; thus two such bases of the respective for-
mule—

may be anticipated, and there is no reason why eurhohine should not
be the amido-derivative of the second of these instead of that of the
first.

The following experiments which were undertaken with a hope to
answer this question have not definitely settled it, but they have
resulted in the discovery of the second isomeric quinoxaline and are
therefore worthy of being communicated.

Simultaneous Oxidation of a- and B-Naphthol and Orthotoluyléne-
diamine.

If molecular quantities of a-naphthol and orthotoluylenediamine
are exposed in an alkaline aqueous solution to the action of potassic
ferricyanide, a blue precipitate is obtained, which dissolves in alcohol
with a magnificent blue coloration. This dyestuff is very unstable,
undergoing spontaneous decomposition in a very short time. As itis,
according to all its reactions, a member of the indophenol-group, a
description of it is not within the objects of this communication.

Very different results are obtained, however, if B-naphthol be sub-
stituted for the a-derivative in this reaction. In this case a dark
somewhat tarry precipitate is formed which is collected on a filter
and carefully washed with water. It is then treated with dilute
(15 per cent.) hydrochloric acid and boiled with it for some time; this
treatment is repeated as long as the acid acquires a yellow tint. The
dark insoluble substance remaining after this treatment is put
aside, whilst the acid extracts are precipitated with caustic soda
solution. The grey precipitate thus obtained is collected, washed,
dried, and recrystallised from a mixture of glacial acetic acid and
alcohol. By this means interlaced needles of a pale straw colour are
obtained, which for analysis may be recrystallised from boiling
toluene. The substance is thus obtained in the shape of long, well-
defined, glistening needles of remarkably definite melting poiut,

404 WITT: THE EURHODINES,

namely, 179°8°, and on analysis gave results agreeing with the
formula of a naphthylenetoluquinoxaline, C,;H,.N2.

Found.

Calculated ao
for C,,;H),N>. I II.

83-50 —
5°34 —
— 11°15

This substance is distinctly different from Hinsberg’s naphthylene-
toluquinoxaline, which melts at 139—142°, and is far more soluble in
all the ordinary solvents. Both substances dissolve with an intense
red coloration in concentrated sulphuric acid, but the solution of the
new substance is of a much more bluish shade. Heated in a dry
test tube, the new quinoxaline is volatilised without the slightest
decomposition. It dissolves in hydrochloric acid with a bright yellow
coloration, and on adding water to this solution, it deposits the hydro-
chloride in yellow crystals. When treated with fuming sulphuric
acid, it is transformed into a sulphonic derivative, whilst concen-
trated nitric acid gives rise to the formation of a well-defined nitro-
derivative.

In the production of this base, other oxidising agents may be
substituted for potassic ferricyanide. The substance is formed equally
well with hydric peroxide, chloride of lime, or freshly prepared
plumbic or manganic peroxide. Its formation may be explained by
the equation—

C;HyN, + CyH,O + O, = CyH,»N, + 3H,0.

As only two naphthylenetoluquinoxalines are possible, my new
compound must have the constitution—

N
AN/N\/.\N on;

|
WAY

it is a derivative of the missing {8 naphthaquinone, and as may
casually be remarked, the first ff-derivative of naphthalene ever
prepared.

Before concluding I have to explain my reasons for adopting for-
mul for the quinoxalines different from those given to them by
Hinsberg.

Throughout the experiments connected with this research, I could
not help being constantly reminded of two well-known substances,
viz., anthraquinone and acridine. Between these two there exists a

A NEW CLASS OF COLOURING MATTERS. 405

definite relationship expressed by the rational formule now universally
adopted for them.
oO

ASA
QOH
O

This relationship becomes still more evident if the anthracene
formala be added as a connecting link.

Now if my formula for the quinoxalines be adopted, the similarity of
these to anthraquinone and acridine becomes a matter of course,
whilst it remains a mystery if Hinsberg’s formule be adhered to.

8b-Naphthylenetoluquinoxaline.

N
4 yyy}

| |
\\/ — WYYVyy

My formula. Hinsberg’s formula.

"\
N

CH;

Moveover, with Hinsberg’s formule great difficulties arise if an
attempt be made to fill in the double connections of the benzene
nuclei. And, notwithstanding all that has been said against them, we
have the less reason to treat them lightly, since Fittig’s elegant
synthesis of a-naphthol from phenylparaconic acid has actually
demonstrated their presence. Finally, I may plead in favour of my
formule that all the other reactions of the quinosalines, as well as
their mode of formation from fatty diketones, may quite as well be
accounted for by using my formule instead of those of Hinsberg.

I am engaged in the continuation of my researches on the eurho-
dine class of dyes, and I hope to be able to lay before the Society, in
a subsequent commuuication, the results of experiments now in
progress.

Charlottenburg, near Berlin.

XLI.—Parabenzylphenol and its Derivatives (Part III), and on an
Isomeric Benzylphenol.

By Eowarp H. Renniz, M.A. (Sydney), D.Sc. (London), Professor of
Chemistry in the University of Adelaide, S.A.

In my former papers (Trans., 1882, 33, 220), the benzylphenol
discovered by Paterno was shown to belong to the para-series of
compounds, and many of its derivatives were described. Since those
papers were published, want of time and opportunity has prevented
me from continuing the work; some results recently obtained, how-
ever, are now laid before the Society, in the hope that the investigation,
which has again been interrupted, may before long be completed.

In preparing material for the earlier part of this work, an oil was
expressed from the crude crystalline parabenzylphenol; this was
mixed with the alcoholic mother-liquor from the recrystallisation of
the latter, and preserved. An examination of this product has led
to some interesting results.

The alcohol having been evaporated, the residue was mixed with a
suitable proportion of ordinary concentrated sulphuric acid, and
heated for some time over the water-bath. When nearly the whole
had dissolved, the mixture was diluted with water, neutralised with
barium carbonate, well boiled, and filtered. To the filtrate, barium
hydrate was added in slight excess, causing a bulky precipitate of the
basic barium parabenzylphenolsulphonate. This was collected and
washed, and the filtrate reserved for examination.

Further Examination of Dinitroparabenzylphenol.—The precipitate
jast described having been converted in the ordinary way into the
potassium salt, the latter was purified by recrystallisation, and used
as a source of the dinitro-derivative already described (Trans., 1882,
222, 226), in order to examine this substance more closely.

The potassium salt was first converted into the nitrosulphonate by
the method already given, the latter suspended in slightly diluted
nitric acid (through which a current of air had been passed for some
time), and the whole warmed on the water-bath with stirring, until
an oily liquid began to separate. During this operation, a strong
smell of benzaldehyde was noticed. The oil, which solidified on
cooling, was collected, and purified by several recrystallisations. It
was identical in appearance with that furmerly prepared, but the
melting point was 85—86°, a little lower than that previously
observed (87—88°). Some of this substance was converted into the

RENNIE: PARABENZYLPHENOL AND ITS DERIVATIVES. 407

potassium salt, and the latter recrystallised. It proved to be anhydrous,
and gave the following result on analysis :—

0'1118 gram substance gave 0°0315 gram K,SO, = 0°0142 gram
K = 12 63 per cent.

Theory for CsH;-CH,-C.H2(NO,)..OK = 12°50 per cent.

On adding a solution of the potassium salt to a solution of barium
chloride, a dense yellow precipitate was formed. This is sparingly
soluble in boiling water, and crystallises therefrom in yellow anhydrous
prisms. On analysis—

0°2545 gram substance yielded 0-086 gram BaSO, = 0°05056 gram
Ba = 19°86 per cent.

C H,°CH.°C,H, NO. 20
Theory for Ci CH-CH(NO}'0> = 20-06 per cent.

On warming the dinitro-derivative gently with concentrated nitric
acid, it was converted, as in previous experiments, into the trinitro.
derivative; but on boiling with slightly diluted nitric acid, red fumes
came off, a strong smell of benzaldehyde was noticed, and the sub-
stance entirely disappeared, no precipitate being formed on the
addition of water. On evaporating repeatedly with water to get rid
of the excess of nitric acid, and cooling, a pale yellow substance
crystallised out; this after two or three crystallisations from water
melted at 121—122°. When boiled with potassic carbonate and
cooled, it yielded the characteristic needles of potassium trinitro-

phenol. On analysis—

0°1623 substance gave 0°053 gram K,SO, = 0°2375 gram K = 1-462

per cent.
Theory for C,H.(NO.);0K = 14°60 per cent.

This decomposition, before unnoticed, takes place easily, and
explains the small yield of the dinitro-derivative referred to in a
previous paper: in fact, as above mentioned, there was a strong smell
of benzaldehyde noticeable in the preparation of this derivative
(doubtless from the oxidation of the benzyl-group), and the liquid
when filtered, evaporated, and boiled with potassic carbonate, yielded
a large crop of crystals of potassium trinitrophenol. Crystals obtained
in this way gave the following results on analysis :—

0°508 gram substance gave 0168 gram K,SO, = 0°07531 gram
K = 14°82 per cent.
Theory 14°60 per cent.

The formation of trinitrophenol in this way (precisely analogous to
the formation of ortho-brom-a-dinitrophenol [C,H,(OH)NO,NO,Br =

408 RENNIE: PARABENZYLPHENOL AND ITS DERIVATIVES,

1:2:4:6] from nitro-bromobenzylphenol) shows clearly that in
dinitrobenzylphenol the nitro-groups occupy the two ortho-positions.

Examination of the Filtrate after Precipitation of the Basic Barium
Parabenzylphenolsulphonate——The filtrate after removal of the basic
barium salt was evaporated to the consistency of a syrup, but as it
did not promise to crystallise well, potassium carbonate was added in
slight excess, the barium carbonate filtered off, and the filtrate evapo-
rated. On standing, it became a mass of crystals. These were dried
by the filter-pump, and recrystallised until colourless. Prepared in
this way, the salt forms a mass of fine white needles. The following
are the results of analysis :—

0°641 gram air-dried substance lost 0°0855 gram at 100° = 13°34
per cent.

0°475 gram air-dried substance lost 0°0625 gram at 100° = 13°15
per cent.

Theory for C;5H;-CH,°C;H;(OH)-SO,;K + 2}H,0 = 12°97 per cent.

0°5555 gram anhydrous salt gave 0°161 gram K,SO, = 0°722 gram
K = 12°99 per cent.
Theory for C;H;-CH,°C,H;(OH)SO;K = 12°91 per cent.

The salt, therefore, is a potassium benzylphenolsulphonate, but it
differs from the corresponding salt of parabenzylphenol entirely in
appearance, and in containing 2} mols. H,0; it must be derived,
therefore, either from an ortho- or a meta-benzylphenol.

Action of Nitric Acid on the new Sulphonate.—On mixing this salt
with dilute nitric acid (1:1), it is easily converted into a nitro-
sulphonate, sparingly soluble in water and alcohol, and crystallising
from the latter in pale-yellow shining anhydrous scales, which in a
thin layer resemble gold leaf. On analysis—

0°351 gram substance gave 0°0875 gram K,SO, = 0°0392 gram

K = 11°17 per cent.
Theory for C,H;CH,°C,H,,NO,OH'SO,K = 11°24 per cent.

On warming this nitrosulphonate on the water-bath with slightly
diluted nitric acid, an oily substance soon separated, which solidified
to a crystalline mass on cooling. After repeated crystallisations from
alcohol, it formed pale yellow rosettes, melting at 81—82°. It is
much paler in colour than dinitroparabenzylphenol. On analysis, it
yielded the following result :—

0-208 gram substance yielded 19°6 c.c. nitrogen at 28°, and 756 mm.
= 0°17 cc. at 0°, and 760 mm. = 0°022125 gram = 1063 per
cent.

Theory requires 10°22 per cent.

AND ON AN ISOMERIC BENZYLPHENOL. 409

It was converted into the potassium salt in the usual way and
yielded pale orange needles, totally different in colour from the
corresponding compound derived from parabenzylphenol. After
recrystallisation, the following results were obtained on analysis :—

Three determinations of water of crystallisation gave 5°45, 5°66,
and 5°40 per cent. respectively, 1 mol. H,O requiring 5-45 per cent.

Two determinations of potassium gave 12°35 and 12°32 per cent.
respectively. Theory requiring 12°50 per cent.

With barium chloride, a solution of the potassium salt gives a
bulky yellow precipitate, sparingly soluble in boiling water, and
crystallising therefrom in dark yellow needles. On analysis—

0:466 gram substance (air-dried) lost 0°012 gram at 100° = 2°57
per cent.
Theory for 1 mol. H,O requires 2°56 per cent., leaving 0-454
gram anhydrous salt, which yielded—

0°1504 gram BaSO, = 0°08843 gram Ba = 19°5 per cent.
Theory = 20°06 per cent.

In the anhydrous condition, this salt is deep red, but on exposure to
the air it speedily absorbs water, and regains its original colour and
weight. The substance from which these salts are derived is

evidently a dinitrobenzylphenol, but it differs widely from its iso-
meride previously described.

During its preparation, a slight smell of benzaldehyde is developed,
but on boiling the pure substance for some time with slightly diluted
nitric acid, no appreciable quantity of trinitrophenol is produced.
Other decomposition products seem to be formed, the nature of which
remains undetermined for the present. When treated with a mix-
ture of concentrated nitric and sulphuric acids, it slowly dissolves ; and
on adding water a substance is precipitated, which can be obtained
by crystallisation from alcohol, in pale lemon-yellow needles. It is
but sparingly soluble in cold alcohol, and yields a yellow potassium
salt; it is probably a higher nitro-derivative, but at present material
is wanting to decide the question.

Action of Bromine on the new Sulphonate.—Six grams of the potas-
sium salt were dissolved in water, and about 3°5 grams of bromine
dropped in slowly, with constant agitation. The mixture became
slightly warm, and deposited white flocks of a substance which
melted to an oily liquid on heating to boiling. As the liquor cooled,
the oil solidified to a mass of crystals, leaving the supernatant fluid
clear. The experiment was repeated with the addition of ice, but it
was not found possible to prevent the formation of the white crystal-
line substance. The latter is easily soluble in alcohol, ether, acetic

VOL. XLIX, 2F

410 RENNIE: PARABENZYLPHENOL AND ITS DERIVATIVES.

acid, and solution of the caustic alkalis, but is practically insoluble
in water. It crystallises from a mixture of alcohol and water in a
mass of minute silky needles, which melt about 91°. It is probably a
dibromo-derivative, but at present material is wanted for further
examination. When warmed with potassium dichromate and sul-
phuric acid, it yields a yellowish substance, probably a quinone; this
awaits further investigation.

The clear aqueous liquid from which the substance just described
had separated, yielded a crop of crystals on evaporation. After
recrystallisation, these form small anhydrous prisms, having the
peculiar white glistening appearance of the phenolbromosulphonates.
On analysis they gave the following results :—

0442 gram substance gave 0°100 gram K,SO, = 004482 K =

10°14 per cent.

Theory for C,H;;CH,"C,H,(OH)Br'SO;K = 10°23 per cent.

An attempt was made to prepare a nitrobromo-derivative by the
action of nitric acid on this salt, but apparently bromine was given off
and a nitrosulphonate formed.

Action of Bromine on the new Nitrosulphonate——When a solution of
the potassium nitrosulphonate is shaken with the requisite quantity of
bromine (slowly added), the latter slowly disappears, and a yellowish
substance separates, which can be obtained in straw-coloured scales
by crystallisation from alcohol. The specimen thus prepared began
to show signs of melting at 105°, but did not become quite liquid
until 109—110°. It was small in quantity, and, perhaps, was not
thoroughly purified. A bromine determination gave the following

results :—

0°1095 gram substance gave 0°066 gram AgBr = 0°02808 gram
Br = 26°51 per cent.
Theory requires 25°97 per cent.

On boiling with potassic carbonate, it yielded a red potassium-
derivative, which did not crystallise easily, being rather soluble in
water. The quantity was so small that it could not be thoroughly
purified; the numbers obtained on analysis, however, leave little
doubt that the substance is a nitro-bromo-benzylphenol isomeric
with that previously described. One specimen of the potassium salt
yielded 10°62 per cent. of potassium instead of 11°27 per cent. as
required by theory.

All the compounds described in the latter part of this paper differ
so widely from the derivatives of parabenzylphenol, that there can be
no doubt that they are derived either from ortho- or from metabenzyl-
phenol, probably the former; at present, however, there is no evidence

PICKERING ON WATER OF ORYSTALLISATION. 411

in favour of one or the other. I hope shortly to obtain more material
and continue this investigation with a view of examining these new
compounds more closely; also to ascertain if possible whether the
higher portions of the distillate from which the crude benzylphenol
is separated contains any dibenzylphenol, a substance yet unknown.

XLII.—On Water of Crystallisation.

By SrenceR Umrrevitte Picxerine, M.A., Professor of Chemistry at
Bedford College.

Tue nature of water of crystallisation, and the relation which it bears
to the salt with which it is associated, has always been, and still
remains, one of the problems least understood in the domain of
chemistry. The energy with which it is held by many salts, the
thermal phenomena attending its combination, and the marked changes
which its presence induces in the appearance and physical character-
istics of a salt, would lead us to rank these hydrates amongst true
chemical compounds: on the other hand, the fact that it may
generally be removed without involving further decomposition of the
molecule, that it does not, in the few cases where investigation has
been possible, alter the chemical reactions of the salt, and further,
that its presence can rarely if ever be expressed according to our
ordinary ideas of atomic valency, must ever differentiate hydrated
salts from “atomic ”’ compounds, although that difference may be one
of degree only and not of kind.*

This difference would appear to have given rise to various unwar-
rantable assumptions as to the nature of these compounds, and to
have led in some cases to the ignoring altogether the possibility of
their being subject to many of the laws which, with atomic com-
pounds, are of fandamental importance.

Till the true nature of hydrates be established this possibility must

* In a paper read before this Society (Abst. of Proc., 1885, 122, and pamphlet),
I suggested that these “molecular” or “ residual” compounds owed their exist-
ence to the valencies of the atoms not being representable by whole numbers, thus
in any molecule a certain amount of unsaturated or residual valency would exist
capable of exerting itself on another molecule, and thereby combining with
it to form a so-called molecular compound. Such compounds, therefore, may still
be regarded as being composed of integral molecules, although their existence will
be due to no new “molecular attraction,” but to the known attraction of atoms
only,

2Fr2

412 PICKERING ON WATER OF ORYSTALLISATION.

never be overlooked, and except on the strongest evidence no argu-
ments must be admitted but such as would hold good if applied to
atomic compounds.

In the present communication, the validity of two very important
conclusions which have been drawn respecting water of crystallisation,
and which would appear to be universally accepted, will be discussed.

These conclusions are :—

(1.) That the various molecules of water in a hydrated salt are not
always similar in every respect, but that they differ from each
other in their heat of combination, their specific volume, and
in other points; that, in fact, a hydrated salt is not always, as
regards its water, a symmetrical substance.

(2.) That in some cases one or more of these water molecules is of
a totally different nature to the others, being water of constitu-
tion and not water of crystallisation at all.

The term “water of constitution” or of “ halhydration” was
originally applied by Graham (Trans. Roy. Soc. Edin., 13, 297; Chem.
Soc. Mem.,1, 82,106 ; Phil. Mag., 6,10, 11) to one of the molecules of
water in the sulphates of the magnesian metals: unfortunately, how-

ever, it is a misnomer, and as such it has created a misconception of
the views which it represents (see p. 429).

The gist of this theory is that the so-called constitutional water is
not water at all, but hydrogen and oxygen forming part of the salt
nucleus itself; crystallised magnesium sulphate is not the hepta-
hydrate of SO.(Mg0O,) but the hexhydrate of SO(MgO,)(HO)., that
this latter salt (improperly regarded as a monohydrate, SO.(MgO,.)H,0)
is an “atomic” compound decomposable by heat into SO,(MgQ,.) +
H,0, just as monosodium phosphate, PO(NaO)(HO)., is decomposed
by the same agent into a totally different salt, PO,(NaO) and H,0.

This theory has been further extended towards a second water
molecule in the sulphates, and is generally expressed in terms
which admit of no misconstruction, thus* :—SO,(OH).,H.O “may be
regarded as a tetrabasic acid of the formula SO(OH),.

Salts of this tetrabasic acid are known. . . . A third hydrate,
SO.(OH),,2H,0, corresponding to a hexabasic acid, S(OH),, was ob-
tained. . . . Salts of this hexabasic acid are also known.
Sulphuric acid forms several classes of salts, of which the following
compounds may be taken as typical examples :—Tetrabasic zincic
sulphate, SO(ZnO,).+ ; hexabasic zincic sulphate, S(ZnO,);; crystal-
lised gypsum, S(CaO,)(HO),; gypsum dried at 100°, SO(CaO,)(HO),;
gypsum dried at 200°, SO,(CaO,).”

* Frankland and Japp (Znorg. Chem., 292—295).

+ Graphic formule being given in the text quoted.

PICKERING ON WATER OF CRYSTALLISATION,

A. General Considerations.

It will be seen at the outset that the representation of the water
present in some salts as being “ constitutional,” increases rather than
decreases our difficulties in explaining so-called water of crystallisa-
tion, for a vast number of salts contain a much larger proportion of
water than can possibly be represented as constitutional, and we are
consequently forced to class them in a different category from others,
which according to their properties and reactions should be precisely
similar compounds.

The constitutional water theory would seem to borrow an air of
probability from a study of the so-called organic acids. The water
with which these are united is in no case greater than that which
would be necessary to form the “ ortho-” acid, and the existence of
perfectly definite ethereal salts corresponding to these acids, and of
similar substances derived from the aldehydes, renders it most pro-
bable that these hydrated acids are in reality ‘“‘ ortho-” acids or acid
hydrols and not mere hydrates. But it must be remembered at the
same time that if such is their constitution they are not real acids at
all but alcohols, and that though they may form “ethereal salts” with
organic radicles, we should not, therefore, expect them necessarily to
form true salts with the metals. Nor do they appear todo so. A study
of the better known salts, the formates, acetates, carbonates, and
oxalates for instance, shows that the proportion of water in the hy-
drated salts, or of metallic oxide in the basic salts, is continually in
excess of that which it is possible to represent as atomically combined,
and that no greater number of instances exist where the proportion
would permit of the salt being regarded as a derivative of the ortho-
acid, than where the proportion would not do so. There are a few
cases* where it is usual to regard a salt as a derivative of the “ ortho-”
acid, but whether they are such is merely a matter of speculation
unsupported by any evidence whatever.

With inorganic acids, the case would appear to be different. These
acids do not affect the “‘ ortho” form more readily than their salts do.
There is but one instance, Sb(OH),, where an ortho-acid is known
without having salts corresponding to it; and again, some of the acids
form compounds with water, which must undoubtedly be regarded as

* Such as the acid formates, H‘COONa, H-COOH, H,O being considered as
HC(ONa)(OH)-O-CH(OH)., although the water present in them is a very variable
quantity and far below the amount required by the formula (Bineau, Annalen [3],
19, 291; 21, 183); also malachite, Cu,CO,,H,0, and azurite,Cu;H,(CO,).; CaCO,,H,0,
and one or two others, though it is difficult to believe that these basic salts are
totally different in their constitution from the other basic carbonates of copper,
cadmium, cobalt, lead, nickel, and zine, which contain too large a proportion of the
metal to admit of such a method of representation.

414 PICKERING ON WATER OF ORYSTALLISATION.

hydrates merely, such as 2H;AsO,,H,O and H,SO;,8H,0, both known in
the solid condition, while Berthelot’s researches on the heat produced
by dilution,* have shown that in the dissolved state hydrates exist cor-
responding approximately to the formule HNO;,2H,0 ; HNO,,15H,0 ;
HCl1,6°5H,0 ; HBr,4°5H,0; and HI,4°5H,0.

It cannot, of course, for a moment be doubted that many metallic
salts do exist containing hydrogen and oxygen as part of the salt
molecule. It will, therefore, be necessary to examine the reasons
which have led to the conclusion in such cases, in order to
ascertain whether similar reasons hold good in other cases under
dispute, and especially in the case of the sulphates.

These reasons may be summed up as follows :—

(1.) The existence of a large number of reactions which are
most consistently represented by regarding the salt as contain-
ing a certain radicle such as NO, in the nitrates, PO in the
phosphates, &c. These reactions certainly point to SO, and
not S being the radicle in the sulphates, the acid acting as a
dibasic, and not as a tetrabasic or hexabasic one.

(2.) The salt cannot be regarded as containing water unless its
formula is doubled, whereas its formation and reactions are
opposed to such a course. The phosphites afford an instance
in point. This argument, however, cannot be applied to the
sulphates.

(3.) Where the hydrogen in a salt is displaceable by a metal, as it
is in the supposed orthosulphuric acid, we find that—

(a.) If in considerable quantity, the salt is strongly acid.

(b.) A metal may be substituted for it atom by atom.

(c.) A large number of salts exist in which the whole of it
is displaced by a metal.

(a.) Now, if a hydrated sulphate be a derivative of orthosulphuric
acid, it will contain no less than four atoms of acid hydrogen; the
presence of some acid hydrogen, it is true, does not necessarily
produce an acid reaction, but it is reasonable to expect that a com-
pound with sucha large amount of it would be strongly acid, instead of
being, as it is in every case where partial dissociation has not occurred,
absolutely neutral. The phosphates, NaH,PQ,, with a much smaller
proportion of acid hydrogen, are strongly acid.

* Annalen [5], 4,446; Compt. rend., 86,279. Berthelot discusses the conclu-
sions drawn by Roscoe and Dittmar (Chem. Soc. J., 12, 128; 18, 156) on the com-
position of the acids when boiling under various pressures, and shows that they are
consistent with the idea that the solutions boiling at 760 mm. consist of definite

hough not very stable hydrates (Méc. Chem., 1, 517 ; 2, 149).

PICKERING ON WATER OF ORYSTALLISATION. 415

(b.) The second feature is a more important one. In no case with
which we are acquainted, except that of metastannic acid,* do we
find the acid hydrogen displaced only in stages of two atoms at a
time, the acid salts contain either an even or uneven number ot
hydrogen-atoms indifferently. With sulphuric acid itself, this is the
case as regards two of the hydrogen-atoms, we get NaHSO, as well
as Na,SO,, but if this same acid is to be regarded as containing six
such hydrogen-atoms, then the remaining four are certainly different
from the first two, for (regarding basic salts for a moment as true
salts) when displaced at all, they are always displaced in pairs.
Surely this must show that they are present in pairs, as in H,O, and
not as separate independent atoms.

(c.) The third point is still more important. Instead of there
existing a large number of metallic orthosulphates, there are none.
Basic salts, I maintain, cannot be regarded as neutral salts. The
method of their formation and their characteristics are entirely
opposed to such a view; but, besides this, we have two very powerful
arguments against such a supposition: (1) The very metals which
displace acid hydrogen most energetically—the alkali metals—are the
very ones which form no basic salts, and, generally speaking, the
greater the energy with which a metal displaces this hydrogen, and
the more stable the salt is which it thereby forms, the less tendency
does it show to form a basic salt. (2) By far the greater number of
the basic salts contain too large a proportion of oxygen to be repre-
sented as ortho-salts. So many of those which are said to exist are
indefinite mixtures, that I will mention those only which I have
had occasion to investigate myself, the sulphates of copper, iron,
and aluminium (Trans., 1880, 807; Chem. News, 45, 121, 133,
146; 47, 181). The only definite compounds were found to be
CuSO,,2Cx0, CuSO,3Cu0, Fe,(SO,);5H,0.¢ Only the first of
these contains a proportion of metallic oxide which would permit of
its being considered as an ortho-salt (and even here the water with
which it is combined would indicate that the formula should be
doubled, see Chem. News, 52, 2), and it is scarcely rational to consider
it as an atomic compound, while the other similar basic copper salt and
the basic iron salt, both of which are more definite and more stable
than the former, must perforce be regarded as mere molecular com-
pounds. That a basic salt is in every respect analogous to a hydrated
salt is the most rational view to take of its constitution, but such a
view is possible only when the water in the hydrated salt is regarded
as such, and not as hydrogen and oxygen.

A study of the basic salts which an acid forms is, I believe, one of

* Possibly also the polymetaborates, but this is by no means certain.
t Three compounds only where thirty were said to exist.

416 PICKERING ON WATER OF CRYSTALLISATION.

the most valuable means of ascertaining the nature of the acid.
Many of the reasons which would lead us to deny the existence of
orthosulphuric acid might be urged with equal force against that of
orthoboric acid, but the manner in which the neutral orthoborates are
obtained, and the character of the metals contained in them, shows at
once that they are true salts (see Bloxam, Chem. Soc. J., 12, 177; 14,
143).
(4.) When a true acid ortho-salt is heated so as to be decomposed
with evolution of water, the residue, the meta- or pyro-
salt, exhibits characteristic reactions which show it to be a
different compound from the ortho-salt. This can only be veri-
fied in cases where the distinction between the salts is further
characterised by the lower ones (meta or pyro) not combining
with water except under special conditions, or after the lapse
of an appreciable, often a very long, time. The phosphates,
antimonates, and stannates form instances in point; with the
arsenates and borates, however, mere contact with water is
sufficient to convert them into the ortho-salts, and, therefore,
no special force can be given to the argument that the sulphates
yield identical solutions whatever the state of hydration in
which the salt may have been before its dissolution.

Only one other plea might be urged for the existence of ortho-
sulphates if it appeared that the majority, or even some noticeable
proportion of hydrated salts contained 2H,0, or of basic salts, 2M’’0.
The answer to this is singularly clear. Of all the hydrates ordinarily
met with, there are fewer with two molecules of water than with any
other number; in fact, but one dihydrate* is known, CaSO,,2H,0,
and even here a considerable weight of evidence derived from various
sources proves that the molecule is more complicated, and that it
should be doubled, the water being removable in fourths. Basic
sulphates with 2M”O are scarcely more numerous than dihydrates.
Those of Cu, Di, G, Hg, and Yt have been described, but what weight
should be attached to the statements as to their existence may be
judged from what has already been said respecting basic salts.

It is not possible, even in the case of sulphuric acid itself, to argue ,

* By partial dehydration and special methods, dihydrated sulphates such as
those of Cd, Cu, Mg, Za, Co, Mn, and Fe may be prepared, but they cannot be
described as salts in a normal state of hydration, and indeed doubts must in many
cases be entertained as to the definiteness of their nature.

+ The introduction of ZnSO,,2ZnO in the passage quoted on p. 412 as an instance
of an orthosulphate, is even less happy than that of gypsum. I have failed to find
any mention whatever of such a compound. It may also be remarked that not the
slightest indication of the hydrate CaSO,,H,O has ever been obtained ; the salt
dried at 100—120° has the composition represented by 2CaSO,,H,0.

PICKERING ON WATER OF CRYSTALLISATION. 417

that the compounds which it forms with water are not mere hydrates
There is a considerable development of heat when the highest isolated
compound, H,S0Q,,2H,0, is dissolved in water, indicating, no doubt,
the existence of still higher hydrates, and there is no sudden break
which would suggest that these are different in nature from the lower
ones; they might all, even including SO,;,H,0 itself, well be members
of the same series ;* the numbers are approximately—

SO, + H,O = 21,320 cal.
H.SO, + H,O = 6,379
H,SO,,H,0 + H,O = 3,039
H,S0,2H,0 + H,O = 1,719
H,S0,,3H,0 + H,O = 1,183
H,S0,,4H,0 + H,O = 750
H,S0,,5H,0 + H,O= 550

Enough will have been said to render any further discussion as to
the existence of two constitutional water molecules in the sulphates
superfluous. All the arguments brought forward against these two
molecules apply with equal force against the existence of only one,
but in opposition to these arguments we find various reasons given in
favour of there being one such molecule present. These reasons may
be included under the title of—

B. Experimental Evidence.

Three considerations appear to have conduced originally to the

’ molecule in the

theory of there being one ‘constitutional water’
magnesian sulphates. These were—-

(1.) That considerably more heat is developed in the combination
of one molecule of water with the anhydrous salt than in the
subsequent addition of any other single molecule.

(2.) That one of the molecules of water present is retained at a
much higher temperature than any of the others.

(3.) That one only of them can be displaced by a molecule of
sulphate of potassium or some allied metal, forming a double salt

such as MgSO,,K.S0,,6H,0.

Heat of Hydration.

The general conclusions deduced from experiments on the heat of
hydration of salts is that, in most cases, the various water molecules
in a hydrated salt bear different relations to the salt nucleus. It will
be found, however, that these deductions rest on a most unwarrantable

* The eame equation will represent all the results, within the limit of experi-
mental error (Thomsen, Thermochem., 8, 57).

418 PICKERING ON WATER OF CRYSTALLISATION.

assumption, namely, that the relation which the water molecules bear
towards each other and towards the salt nucleus is absolutely unalter-
able by any process short of their removal.

We are indebted to Thomsen (J. pr. Chem., 18, 5) for the most
complete series of investigations on this subject. It is true that I
have already (Trans., 1885, 102) had occasion to show that Thomsen’s
data were incorrect, since the substances with which he dealt were
not definite hydrates but mixtures of different hydrates; but even if
this had been otherwise he would, no doubt, have obtained results of
the same general character, and would have drawn the same general
conclusions from them.

Thomsen attempted to ascertain the heat of combination of the
individual molecules of water in certain hydrated salts by determining
the different heats of dissolution of the same salt hydrated to various
extents. The difference in the heats of dissolution of any two
hydrates, such as MgSQ,,3H,O and MgSO,,4H,0, gave the thermal
value of the reaction [MgSQ,,3H.0 + H,0]. Now this, he assumes,
gives the heat of combination of the fourth molecule of water in the
hydrate MgSO,,4H,0, but unless the above-mentioned unwarrantable
hypothesis be accepted, it does, I maintain, do nothing of the sort.
It can tell us the heat of combination of the fourth molecule only in
the event of the heat of combination of the other three remaining
entirely unaltered by the addition of this fourth molecule, and we
have no more reason for making such an assumption in the case of
these molecular compounds than we have in the case of atomic com-
pounds, and certainly no one would venture to state that the heat of
combination of Pb + S, and the relations which exist between these
atoms in the compound PbS, remain entirely unaltered by their assump-
tion of 4 atoms of oxygen to form PbSOQ,. Lead sulphate does not
contain lead sulphide, no more does a tetrahydrated salt contain the
trihydrate. The thermal reaction [MgS0O,3H,0 + H,O] is made up
of two distinct quantities—(1) the decomposition of the trihydrate,
(2) the formation of the tetrahydrate; and all we can measure is the
algebraic difference between the two. The water molecules in the
one hydrate may be in a state of combination very different from that
in which they exist in another hydrate, and in any one hydrate all the
water molecules may or may not be similar; but experiments such as
Thomsen’s can throw no light whatever on the question.*

* We may push this line of argument still farther, and question, for instance,
whether a determination of the heat evolved in the reaction (MgSO, + H,O) gives
us any information as to the heat of combination of MgSO, with the H,0, or
whether the relations of the atoms composing.the anhydrous salt are not modified
by the combination of the salt with another molecule. I have elsewhere (Chem.
Soc. Proc., 1885, 122) stated that various experimental evidence favours such a

PICKERING ON WATER OF CRYSTALLISATION. 419

The heat which is evolved on the addition of one molecule of water
to the anhydrous salt is certainly very much greater proportionately
than that evolved on the addition of more than one; but even if
Thomsen’s numbers be accepted, they only show that this difference is
one of degree only and not of kind; the first molecule in MgSO,,7H,O
develops in combining 7980 cal., according to him, and the other six
only 17,100 cal. altogether, and yet he proves they may all be mem-
bers of the same series.* A little consideration, moreover, will show
that the facts observed here with hydrated salts are but a counterpart
of those shown by atomic compounds. The heat of combination of
one atom of oxygen with two of hydrogen is 69,000 cal., but the com-
bination of a second actually absorbs 22,400 cal. Who would argue
from this difference that in hydrogen dioxide the two atoms of oxygen
are united to the hydrogen with different degrees of energy and
perform entirely different functions? And yet on the strength of a
precisely similar but much smaller difference, it is argued that the
molecules of water in a hydrated salt are different not only in degree
but also in kind. Such an argument is utterly worthless.

There is one conclusion, however, which may be drawn from
Thomsen’s work. As with magnesium sulphate, so with the sulphates
of zinc, copper, manganese, the double sulphates, and other salts, he
found that the diminution in the heat of dissolution was not always
directly proportional to the amount of water present in them; whereas
with the sulphate,t carbonate, and the two phosphates of sodium it

view, but it is probable that the addition of any fresh molecule to a “ molecular”
compound (H,O to MgSO,,3H,O) would affect the relationship of the various
molecules (MgSO, and 3H,0) to a far greater extent than it would the relationship
of the atoms constituting those molecules.

* This relationship is brought out so clearly in the case of magnesium sulphate,
that it would almost carry conviction with it, in spite of the hypothesis on which it
rests, were it not that Thomsen was certainly dealing with indefinite mixtures, and
that his determination of the heat of dissolution of the monohydrate contains an
error of about 1100 cal., a quantity which would entirely destroy these fancied re-
lations (/oc. sup. cit.). It is instructive to compare the results obtained in the case
of magnesium sulphate with those obtained in some other cases where the water
present cannot be regarded as constitutional, and where, nevertheless, the first
molecules appear to be combined with far greater energy than the subsequent ones,
thus :—

In MgS0O,,7H,0, heat of combination of the—

Ist, 2nd, 8rd, 4th, 5th, 6th, 7th H,O.
6980 2250 3600 3210 2230 2110 3700
In SrBr.,6H,O, heat of combination of the—
1st, 2nd, 3rd, 4th, 5th, 6th H,0.
6150 3800 3100 3100 3160 4000

+ His numbers for the heat of dissolution of the anhydrous salt being emended

in accordance with the author’s observations (Trans., 1884, 689).

420 PICKERING ON WATER OF CRYSTALLISATION.

was so. Such a proportionality is, no doubt, due to the partially
dehydrated specimens being mixtures of the fully hydrated and anhy-
drous salt in various proportions; where, then, this proportionality
does not exist, we may conclude that the partially dehydrated salt
contains some intermediate hydrates, which under favourable circum-
stances might have been isolated, and each of which possesses a special
heat of dissolution different from that of a mixture of the same per-
centage composition.

Temperature of Dehydration.

It is difficult to attach much weight to arguments deduced from any
difference in the temperatures at which the various molecules of
water are expelled from a hydrated salt, for in some cases (that of the
typical magnesium sulphate itself) this difference is very small, not
more than 10° C. (Trans., 1885, 101), and if we admit such reason-
ing here we should have, in order to be consistent, to admit it
in all cases, and devise some different principle of combination for
every molecule of water which is driven off at a different temperature.
But, apart from this consideration, any argument based on such
behaviour will be found, as in the case of Thomsen’s deductions, to be
drawn in oppusition to one of the fundamental principles of chemistry:
that no chemical action is a single action, but is the result of at least
two or three distinct actions. We cannot assert that magnesium
sulphate, MgSO,,7H,0, on being heated gives up six molecules of water
because they are loosely combined, and retains the seventh because it
is more energetically combined; all we can state is that the hepta-
hydrate, being incapable of existing in dry air at 155°, is decomposed,
and the monohydrate, which is capable of existing under such condi-
tions, is formed. These two views of the same action are radically
different: the second is a mere statement of the facts themselves,
whereas the first asserts that a lower hydrate is actually contained in
a higher one—that the monohydrate is a residue from the heptahydrate.
What reasons have we for stating this to be so here any more than in
other cases of decomposition ? Because hydrogen dioxide gives up one
of its atoms of oxygen at a temperature many hundred degrees lower
than that at which the second one is parted with, no one has ever
ventured to suggest that the two oxygen-atoms in hydrogen dioxide
are, therefore, different, one being combined very energetically, the
other very feebly. An argument which will not hold good in one case
must not be blindly accepted in the other. Hydrogen dioxide does
not contain water, and heptahydrated magnesium sulphate does not
contain the monohydrate.

Whatever experimental evidence can be brought to bear on the

PICKERING ON WATER OF CRYSTALLISATION. 421

question certainly tends to show that this is the case; that the further
addition of water to a hydrated salt necessitates an alteration in the
relations existing between the salt nucleus and the water molecule
already combined with it, and that similar intramolecular changes
occur in other cases where so-called molecular compounds are con-
sidered.

(1.) If a fully hydrated salt really contained the lower hydrates, it
is difficult to understand why these Jatter should not be successively
formed in the process of dehydration, since the abstraction of one mole-
cule of water would not, ex hypothesi, influence the stability of the re-
maining ones. But, as is well known, it is generally impossible to
obtain intermediate hydrates by dehydration, even when these are
capable of a separate existence.

(2.) According to Hannay (Chem. Soc. J., 1877, ii, 389), when hydrated
zinc sulphate is heated at 100°, it does not fuse, but as soon as three
molecules of its water have been driven off, it does so. Now the only
probable explanation of a loss of water causing an increase in fusibility
must be that the loss is accompanied by some change in the consti-
tution of the salt.

(3.) If partial dehydration consists of the removal of the more
weakly combined molecules leaving a residue containing the more
energetically combined ones, hydration would exhibit the reverse
phenomena; the lower hydrates containing the strongly combined
molecules would be formed first, and then the higher hydrates.
But this is not the case. I have mentioned elsewhere (Trans.,
1885, 108) that the peculiarities in the manner in which anhydrous
and monohydrated magnesium sulphates dissolve show that the
latter is not formed during the hydration of the former; the anhy-
drous salt cakes but dissolves quickly ; the monohydrate does not
cake, but takes a considerable time to dissolve, rendering the liquid
milky meanwhile. On the view here advocated, no difficulty would
be met in explaining this fact; whichever hydrate was the most
stable under the given conditions, would be formed directly ; none of
the water molecules in it being combined with greater energy than
the others, none of them would be taken up before the others.
What also can be the explanation of the monohydrate taking so
much longer to assimilate six molecules of water than the anhydrous
salt does to assimilate seven, unless it be that some alteration in the
constitution of the salt is necessitated ?

(4.) Hannay (Trans., 1879, 456) mentions another instance, which,
if verified, would go far to show that the addition of a fresh
molecule to a compound molecular group influences the energy with
which the original components of this group were held together.
He states that the double salt, [(ZnSO,,7H,O)(MgS0Q,,7H,0)], when

422 PICKERING ON WATER OF CRYSTALLISATION.

dried at 100°, loses its water at twice the rate with which the com-
ponents, ZnSO,,7H,0 and MgSO,,7H,0, do, and leaves a residue
retaining only 2H,0, whereas a mixture of the uncombined molecules
would retain double this quantity of water.* We cannot, however,
go so far as Hannay does, and say that “the two salts in combining
have expended one-half of the affinity of the anhydrous salt for its
water of crystallisation,” unless it can be proved that “affinity is
inversely proportional to the rate of loss.”

(5.) Some facts which I have recently brought before this Society
(Trans., 1886, 260) afford very strong evidence, though not amounting
to absolute proof, in favour of the views here advocated. It was
ascertained that as the temperature rises, a saline solution expe-
riences changes which would indicate (in most cases) the forma-
tion of higher hydrates; it was also proved that the heat of com-
bination of a salt with its water of crystallisation was not a constant
quantity; and further, that the temperature at which this latter
experienced a diminution was the same at which a higher hydrate
began to form in the solution, as if the energy of combination
between the salt and its water increased as the temperature rose, till
this increase became so considerable that a larger amount of water
became attached, but that the increased energy expended on this
necessitated a diminution of that expended on the water previously
combined. A similar intramolecular compensation was observed in
the case of the double salts; those temperatures at which the heat of
combination of a salt, such as CuSO,,K,S0O,, with its water of crys-
tallisation was greatest, corresponded to those at which the heat of
combination of the component molecules, CuSO, and K,SQ, in the
anhydrous salt, was smallest.

These arguments, meagre though they be, all tend to throw doubt
on a theory which at the best is entirely unsupported by expe-
rimental evidence, and in my opinion is irrational.

The temperature of 100° does not appear to be critical as regards
water of crystallisation. On the one hand there are many instances
in which “ constitutional’ water is evolved at or below 100°, such as
orthoboric and the ortho-organic acids; and on the other hand water,
which must be regarded as water of crystallisation, is, in very many
cases, retained at a much higher temperature ; thus many chlorides,
MnCl, for instance, retain a notable quantity of water even at 220°;

* This statement (1879, 459) is unfortunately irreconcilable with those in a
previous paper (1877, ii, 388 and 399), “ that the dehydration of magnesium sul-
phate might be pushed till only one molecule was left at 100°,” and that with zinc
sulphate at 100°, “the dehydration becomes very slow till siz molecules have gone,
and there it stops.”

PICKERING ON WATER OF CRYSTALLISATION. 423

so also with many metallic oxides: thus K,0,5H,O* is stable at tem-
peratures below 400°. Indeed, the temperature at which water of
crystallisation is evolved, not only from different salts, but also from
different hydrates of the same salt, as well as its rate of evolution,
would appear to show the same amount of variation as is observed in
the temperature and rate of other chemical decompositions.

Although the lower hydrates of a salt are undoubtedly more
stable than the higher ones (just as H,O is more stable than H,0,),
the results of dehydration afford no grounds for considering that
there is any difference either in kind or in degree, in the manner in
which the various water molecules in any particular hydrate are
combined, nor can the temperature of 100° be regarded as being more
critical with reference to water of crystallisation than 50°, 150°, or
any other temperature.t

Double Salts.

The existence of these salts forms the third reason generally
assigned for regarding the magnesian sulphates as containing a mole-
cule of constitutional water, sometimes termed in consequence water
of halhydration. The view that the double salt, such as

MgS0,,K,S0,,6H,0,
is derived from the corresponding magnesian sulphate,
MgS0O,,H,0,6H,0,

by the displacement of a molecule of water in it by potassium sul-
phate, was controverted by Thomsen, who endeavoured to show that
the 6H.O in the one salt is combined with a different heat evolution
to the 6H,O in the other salt. Even if it be admitted that Thomsen’s
work did measure the heat of chemical combination of these water
molecules, the fact that it is different in the one salt to what it is in
the other is surely not an argument that the one is not a true dis-
placement-product of the other; it would, indeed, be very remarkable
if the displacement of the one molecule of water did not affect the

* Maumené (Bull. Soc. Chim., 44, 578) assigns much more complicated formule
to this and similar hydrates, 9K,0,47H,O, &c.

+ It must be remembered that 100° is a temperature which for convenience sake
is generally employed when examining the behaviour of hydrated salts: no investi-
gation has been undertaken, as far as I am aware, to ascertain whether it is preferable
to any other temperature.

The rate of dehydration of some substances has been studied by Hannay (Joe. eit.)
and by Ramsay (Chem. Soc. J., 1877, ii, 395). Miiller-Erzbach (Ber., 17, 1417;
Ann. Russ. Chem. [2], 28, 607; Chem. Centr., 1885, 470) has examined the vapour-
tensions of various hydrated salts. The tensions became inappreciable after the
loss of amounts of water which were very various in different and even similar
cases, and which did not represent any simple molecular proportions.

424 PICKERING ON WATER OF ORYSTALLISATION.

other molecules present. I have elsewhere (this vol., p. 13) pointed
out that the simple manner in which these double salts are formed
from the magnesian sulphates, and the accurate manner in which the
various characteristics of these latter are reproduced in the double
salt, justify us in regarding these as most evident instances of dis-
placement. The argument based on their existence then is that only
one molecule of water in MgSO,,7H,O being thus displaceable, this
one must be different from the other six. The very fact that one of
these water molecules has been displaced may, however, be the very
reason why the others are not displaceable. With atomic compounds,
we know that displacement of one atom affects the others, rendering
their displacement more easy or more difficult, sometimes even im-
possible ; and any argument which is based on the gratuitous assump-
tion that with hydrates no such intramolecular reaction takes place,
must be valueless. We might as well argue that three of the hydrogen-
atoms in benzene itself were different from the other three, because
three only can be displaced by NO,, as that one of the water molecules
in MgS0O,,7H,0 is different from the rest because one only can be
displaced by the K,SO,.*

Not only can no argument in favour of the constitutional theory be
derived from these double salts, but they even afford a strong argu-
ment against this theory.

If the double salt is derived from M’<0>s0<0 1 6H.0, a sul-

phate with one molecule of “ constitutional” water, its constitution
will be M’<0>SO0<9>S0<0,6H,0.

Now I have shown (this vol., p. 1) that the anhydrous copper potas-
sium sulphate and probably all its congeners exist in three distinct
modifications, differentiated by their colour, specific heat, and heat of
dissolution. The first of these modifications is obtained by the simple
dehydration of the hydrated salt at the lowest possible temperature,
whereas the other two are produced from this one only at considerably
higher temperatures. If, therefore, any of these anhydrous salts
contains the constituent sulphates united in the same manner in which
they are in the hydrated compound, it is the first or a-modification
and not either of the others.

Now in all cases where the specific heat of a hydrated salt has been

* It is not difficult to see the reasons which may render the displacement of
more than one molecule improbable. From the numbers obtained, it would seem
probable that the displacement of one molecule of water by K,SO,, has considerably
increased the heat of combination of the other six; but even a small increase in
their heat of combination would render it impossible to displace them by K,S0O,,
for such a displacement would be accompanied by a decided absorption of heat.

PICKERING ON WATER OF CRYSTALLISATION. 425

determined, it appears that this specific heat is equal or very nearly equal
tothe sum of those of theanhydrous salt and the water present (reckoned
as solid water), the heat of dissolution also of such a salt indicates
that it behaves as if it were a mere mixture of such constituents
(this vol., p. 7); of the three modifications of the double salts,
there is only one which acts in a similar manner, it is the a-salt
only which has a specific heat identical, within experimental error,
with that of a mixture of its constituents (p. 14), and which dissolves
with the same* heat evolution as would such a mixture; this modi-
fication therefore, and consequently the hydrated salt from which it
is so simply obtained, is the only one which must be regarded as pos-
sessing a constitution similar to that of a true hydrated salt, the potas-
sium sulphate in it behaving in the same way as water of crystallisa-
tion in the latter. If, then, it be regarded as being derived from the
single sulphate by simple displacement of water, that water must be
water of crystallisation and not water of constitution.

Specific Heat.

It has been ascertained, as indicated above, that the specific heat of
a hydrated salt is equal to the sum of that of the anhydrous salt and

the water reckoned as solid water, and this will be found to hold as
good in the case of the monohydrated sulphates as in any other case
(see the determinations made by Kopp and by Pape). Much stress,
however, cannot be laid on this similarity, for the numbers obtained
show great discrepancies, and can be regarded as being but approxi-
mations at the best; it would probably require the adoption of far
more delicate methods than those at present known, in order to dis-
tinguish between water of crystallisation and the elements of water,
for while the specific heat of the former is 9°, that of H, + O in solid
compounds would appear to be 8°6°¢ (Kopp, Jahresh., 1864, 43).

Volume of Solid Hydrated Salts.

The specific volume of solid hydrated salts would appear to be
subject to considerable variations. In the case of certain highly
hydrated salts, such as sodium sulphate and phosphate, the volume is

* See p. 288, where a misprint of 600 cal. in the numbers given on p. 7 and 15 is
corrected. We may state this fact in another way: the formation of the a-salt from
its constituents evolves 29 cal., the B-salt 3249 cal., and the y-salt 1331 cal., the one
in which the heat of formation is least, the a-salt, is that in which the constituents
are less energetically united, the 8- and -moditications both being formed with
considerable evolution of heat are the two which, if any, are atomic compounds.

+ With reference to the variations in the specific heat of hydrated salts, see
this vol., p. 285,

VOL, XLIX. 26

426 PICKERING ON WATER OF CRYSTALLISATION.

exactly equal to that of the solid water contained in the salt, whereas
in other cases the volume is equal to the sum of those of the salt and
the solid water present (Playfair and Joule, Chem. Soc. Mem., 2, 401;
3, 57,199; Chem. Soc. J., 1, 121, 129).

The volumes of the so-called magnesian sulphates in various states
of hydration were examined by Thorpe and Watts (Trans., 1880,
102). The general result of this examination was to show that
the volumes of the sulphates of this group of metals (Cu, Mn, Fe,
Ni, Co, Mg, Zn) were practically identical in whatever state of hydra-
tion they were, and that the water molecules in the higher hydrates
occupied a larger volume than those in the lower hydrates. The mean
results were as follows :—

Volume of the salt with—
, ts 2g 3, 4, 5, 6, 7H,0.
448 55°5 68°83 833 987 1129 130 1461

from which they conclude—

Vol. of the Ist, 2nd, 3rd, 4th, 5th, 6th, 7th water molecule.
10°7, 13°3, 14°5, 15°4, 142, 17°1, 16:1,

the volume of one molecule of ice at the same temperature being
about 19°8. These numbers can only be regarded as approximations,
as in many cases it was impossible to obtain the required hydrate,
and in others the definiteness of the product was doubtful. However,
they were all obtained by definite processes and not by partial dehy-
dration, and therefore no objection to the data can be made on that
score, as in the case of Thomsen’s experiments on their heat of dis-
solution. But precisely the same objections can be made to the con-
clusions drawn by Thorpe and Watts as to those drawn by Thomsen;
the two cases are exactly parallel. The fact determined is simply the
volumes of #H,O in the hydrate containing that amount of water;
the difference between this and the volume of that with (z + 1)H,0
does not give us the volume of the (# + 1)th molecule, unless we
assume that the volumes of the x molecules have been entirely un-
affected by the addition of another molecule, and this we have no
right whatever to assume. We may say that the volumes of the
lower hydrates are comparatively less than those of the higher ones,
but we cannot say anything whatever about the volumes of the
different molecules in any particular hydrate, they may all be equal
or all unequal; experiments such as these can throw no light on the
question.

The water in the monohydrate would appear to occupy a much
smaller volume than any of the others, indicating no doubt its superior
stability, but the difference is not such as to warrant any conclusion

PICKERING ON WATER OF CRYSTALLISATION. 427

that it is different in kind from the others. The difference between
it and the second molecule for instance, is 2°6, while a difference
greater than this exists between the second and the sixth or seventh,
and even between the fifth and sixth, where there can be no doubt
about both molecules being water of crystallisation.

Volumes of Salts in Solution.

Enough will have been said already to show that the arguments on
which the “constitutional” molecule theory was originally founded
are utterly fallacious; there is, however, another argument deduced
from the volumes occupied by salts in solution which at first sight
would appear to be a strong one.

Nicol (Phil. Mag., 1884, 179) has shown that in a large number of
cases the change of volume accompanying the displacement of one
radicle by another (whether metal or non-metal) in a salt in solution
is a constant quantity, independent of the nature of the other radicle
present: thus the substitution of Br for Cl increases the volume of a
salt solution (1: 100H,O) by 7°4, whether the metal with which the
halogen is combined be sodium or potassium. He also finds that this
holds good whether the salts are hydrated or anhydrous in the solid
condition. Now the increase in volume produced by substituting
Cl for sate he finds to be 9°2 in most cases, but that with the mag-

nesian sulphates it is only 6°8, and consequently, he argues, we are

not substituting Cl for 50 in these latter, but for something else,

SO,
2°
however, from which these conclusions are drawn appear to be very
insuffcient and capable of other interpretations. They are as
follows :—

namely, for + a molecule of constitutional water. The data,

Difference.
Metal in the salts cl- (SO,)
examined. ie
K 89
Na 9°2
Cd .
> 8:9
L 9°8

9-2

PICKERING ON WATER OF CRYSTALLISATION,

Magnesium Sulphates.

6°4

6°8
The separation of these metals into two groups would seem to have
been effected on somewhat arbitrary principles. According to the
class of metal taken, cadmium has certainly a good claim to be ranked
with the magnesium sulphates, for although (like copper and manga-
nese sulphate) it does not crystallise with 7H,O, it forms double salts
with the sulphates of potassium, sodium, and ammonium, isomor-
phous with the corresponding double salts of copper, cobalt, and
nickel. A distinction based on the numbers obtained would seem to
be more arbitrary still. The difference between two members of the

different groups is as small as 1‘1 in the case of L and Ni, whereas
individual members of the same group show a difference of 09

(K and L), and even as much as 1°7 (Cu and Ni).

More concordant results were obtained when the solutions of the
various sulphates were precipitated by barium chloride, the expansion
which took place in the various cases was as follows :—

the numbers in the last four cases being undoubtedly smaller than in
the first two. It may well be questioned, however, whether these data
are not much too meagre to found any general conclusions upon ; there
are so many points of difference between the alkali metals (represented
here by only two members) and the magnesian metals, that it seems
rash to conclude that these different volume-changes are due to one
peculiarity more than to another. Sulphates belonging to other
classes of metals should certainly be examined also. However, accept-
ing the data as sufficient, the conclusion drawn is that the volume of a
molecule of water of constitution is 24°4, or 6°4 greater than that of
water of solution or water of crystallisation ; but further on, p. 192,
Nicol shows, from a comparison of the sodium phosphates, that “the
elements of water in a molecule occupy Jess volume than water,” the

PICKERING ON WATER OF CRYSTALLISATION. 42

volume being 9°9 in one case and 13°9 in another, according as they
are contained in one molecule of the salt or two. This is entirely
irreconcilable with the previous conclusion; the very meaning of the
‘constitutional molecule ” is that it is not water at all, but the elements
of water, the monohydrated sulphates according to this theory are
identical in constitution with the acid phosphates; if then these
elements of water occupy a much smaller volume than H,O, we have
a perfect proof that no such “elements” or constitutional water is
present in the magnesian sulphates. Even were it thoroughly estab-
lished that these salts do exhibit anomalies in their volumes, and that
these anomalies are due to the water combined with them, this would
do no more than show that the different molecules of water of crystal-
lisation may bear a different relation towards the same salt.*

On the strength of his conclusions that this “‘ constitutional” watcr
occupies a greater volume than water of crystallisation, whereas the
experiments of Thorpe and Watts (see above) on solid hydrated salts
would indicate that it should occupy a smaller volume, Nicol con-
cludes that water of crystallisation, or, as we ought to term it,
combined water, does not exist in solution at all, i.e., that a salt
solution contains the salt in the anhydrous condition.

“At first sight,” he proceeds, “the experiments of Thomsen and
others, on the heat of hydration of various salts, lend much support
to the view that water of crystallisation does exist as such in solution ;
for there is no doubt that while in very many cases a hydrated salt
dissolves in water with absorption of heat, the same salt when dehy-
drated evolves heat on solution. But is it not possible that the act of
solution of a dehydrated salt consists, first, in the taking up of water
to form a hydrate ; and that this hydrate, in dissolving, parts with its
water, which thus becomes indistinguishable from the rest of the
water”? The simple answer to this question is that it is not possible.
Nicol seems to have overlooked the fundamental principle of thermo-
chemistry, a principle which is but a special application of the conser-
vation of energy, that the decomposition of any compound absorbs
precisely the same amount of heat as its formation evolves, that if the
hydrate be first formed and then decomposed the thermal results will
he the same as if it had never been formed at all. That a hydrate is
formed, as Nicol admits, when a dehydrated salt is thrown into water
is unquestioned, it can be seen, felt, removed, and analysed; that its
formation is attended with considerable development of heat is equally
sure, the quantity developed can be measured, and is generally a very
considerable quantity. Even if we can reconcile ourselves to the

* Whatever conclusions may be drawn from this work, it must be remembered
that their bearing on the nature of water in a solid salt is but indirect, since they
apply directly to the dissolved salt only.

a aot. LMI my Bia ae

430 PICKERING ON WATER OF CRYSTALLISATION.

belief that the water which will at one moment combine so energeti-
cally with the salt, will the next decompose the very compound which
it has formed; even if we believe that this salt with such a strong
attraction for water is capable of existing in the presence of water
without combining with it, we are driven to admit that the force
which causes it to do so is not only more powerful than the
chemical attraction between the salt and its water, but that in its
operation it evolves more heat than the chemical combination. This
force Nicol considers to be the resultant of the attraction of (1) salt
molecule for salt molecule, (2) water for water, and (3) water for
salt; the first two act in opposition to the last; it is therefore only the
predominance of this last over the sum of the other two which causes
dissolution. It is scarcely necessary to point out that no amount of
attraction, chemical or otherwise, can of itself produce the least
particle of heat; it is only the actions which result from this attrac-
tion which can do so; according to his theory, then, the motion of
the drawing together of the salt and water molecules, although it does
not draw them into that state of close connection known as chemical
combination, must produce a greater amount of heat than does their
actual combination itself. Surely this is quite inadmissible.

Nicol’s views on the nature of solution are based chiefly on the fact
that the volume of water of crystallisation in the liquid state is indis-
tinguishable from that of ordinary water. But why should we attach
more weight to such a fact with liquid salts than we do with solid
salts? Wemight as well argue that in a solid hydrated salt the water
is not combined with the salt because it possesses a specific heat (and
specific volume too, in many cases) indistinguishable from that of solid
water.

Very many more thermal phenomena besides the one above dis-
cussed are inexplicable except on the hypothesis that hydrates do exist
in solution; whether their existence can give us a full explanation of
all the facts about dissolution or not, is quite another question, and
one which it would be foreign to the subject of the present paper to
discuss.*

Conclusion.

The results of the present examination may be summarised as

follows :—
1. The theory that the sulphates contain either one or two molecules

* For this reason, also, I refrain from discussing the results obtained by various
physicists on the alteration of the freezing point, temperature of maximum density,
and tension of water by the addition of salts thereto—no conclusions adverse to the
opinions expres:ed in the present communication are, I believe, deducible from these
results.

PICKERING ON WATER OF CRYSTALLISATION, 431

of “constitutional” water finds no confirmation whatever in general
considerations as to their formation, properties, composition, and
analogies with other salts.

2. The most decisive evidence asto whether a given salt is.an “‘ortho-”
salt or a basic salt is afforded by the nature of the metals which form
such salts, and by ascertaining whether it displaces the hydrogen-
atoms singly or in pairs. ‘Those metals which displace “acid”
hydrogen most energetically are the ones which exhibit least tendency
to form basic salts.

3. Although the heat of hydration of a salt indicates that the lower
hydrates are generally more stable than the higher ones, it gives no
information whatever as to whether the various water molecules in
any particular hydrate are all similarly combined or not.

4. Nor is any information on this point afforded by the different
temperatures at which the water molecules are evolved.

5. Experimental evidence, no less than general considerations, show
that the water molecules in a lower hydrate are affected by the addi-
tion of more water. The higher hydrate does not contain the lower
one.

6. The displacement of one of the water molecules in the magnesian
sulphates by K,SQ,, affords no grounds for the supposition that the
one displaced differed in any respect from the others. The physical
properties of the resulting double salt afford a strong argument in
favour of the displaced molecule being water of crystallisation only.

7. The specific heats of hydrated salts throw no light on the nature
of the different water molecules.

8. Experiments with the solid hydrated sulphates of the magnesian
metals show that the volume of the water in the lower hydrates is
relatively less than that in the higher hydrates. But the difference is
not such as would indicate any difference in nature of any of the
various hydrates, and certainly affords no information as to the
volumes of the individual water molecules in any particular hydrate.

9. The volume of dissolved salts may show that one molecule of
water in these magnesian sulphates differs from the rest; but the data
available are at present too meagre to warrant any definite conclusions
being drawn. At any rate, whatever the difference may be, it is of
exactly the opposite nature to that which would be exhibited if it were
water of “ constitution.”

10. With the exception of this last-mentioned fact, which can at
present be regarded as a possibility only, we have no grounds for
supposing that the various water molecules in a hydrated salt are not
all combined in a similar manner, and do not all occupy exactly the
same position with regard to the salt nucleus ; that, in short, a hydrated
salt is not a symmetrical substance. At the same time, it must be

432 BROWN ON AN ACETIC FERMENT

admitted that we have no positive evidence to show that it is symme-
trical, and we must, therefore, regard it as such provisionally only, till
some unquestionable evidence in favour of or against this view be
forthcoming.

XLITI.—On an Acetic Ferment which forms Cellulose.

By Aprian J. Brown.

During my work on the chemical actions of Bacterium aceti,
described in a previous paper (this vol., p. 172), I met with the
peculiar acetic ferment commonly known as the “vinegar plant” or
“mother.” This ferment differed so much in appearance from any
form of B. aceti I had noticed, that it seemed probable it was a
distinct organism. In order to ascertain this, and also to enable me
to study its chemical actions, I obtained pure cultivations by a com-
bination of the fractional and dilution methods in the way I described
in my previous paper. The nutrient solutions used for this purpose
were composed of red wine diluted with half its bulk of water, and
rendered acid with 1 per cent. of acetic acid in the form of ordinary
vinegar. This liquid strongly favours the growth of the acetic
ferments, and is at the same time very prejudicia] to the growth of
most other organisms.*

In order to be more certain of the purity of the culture, inocula-
tions of it were made in gelatin and beerwort. In about ten days,
well-defined colonies of the ferment commenced to grow in this solid
nutrient mixture. Many of these colonies were transferred separately
to suitable sterilised liquids, and in every case the characteristic
growth of the “vinegar plant” appeared. There can be no doubt,
therefore, that the cultures thus obtained were quite pure.

A pure cultivation of the “vinegar plant,” when commencing to
grow in a liquid favourable to its free development, is usually first
noticed as a jelly-like translucent mass on the surface of the culture
fluid; this growth rapidly increases until the whole surface of the
liquid is covered with a gelatinous membrane, which, under very
favourable circumstances, may attain a thickness of 25 mm. This
membrane is slightly heavier than water, and when gently agitated

* In all experiments mentioned in this paper. the same methods were used, and
the same precautions taken, with regard to the sterilising of culture fluids, cotton-
wool, &c., as have been previously described by me (this vol., p. 173).

~~ ne m—e Atm aa.

Qo Sow we et i Oe

WHICH FORMS CELLULOSE. 433.

sinks until its upper surface is covered with liquid, when another
layer of the ferment at once commences to grow above the old one.
Frequently as many as five or six layers of growth are thus formed,
causing the whole mass to appear striated when observed laterally.
In colour, the membrane is white and translucent, unless grown in
coloured solutions, in which case it assumes somewhat the colour of
the liquid.

On removing the membrane from the liquid, it is found to be very
tough, especially if an attempt is made to tear it across its plane of
growth ; parallel to this plane, however, it is much more easily divided
into a number of layers, evidently representing successive stages of
growth. In touch and general appearance, the “ vinegar plant” has
a remarkable resemblance to a soft animal membrane.

If the “vinegar plant” is grown in a liquid unfavourable to its
free growth, such as yeast-water, it is first observed as a jelly-like, very
transparent mass at the bottom of the solution; and this gradually
increases in size until the surface is reached, and the liquid appears tu
be almost entirely filled with it. Close examination of this mass,
however, shows that it is composed of extremely attenuated mem-
branes of the ferment, which ramify through the bulk of the liquid,
and actually enclose much of it, thus giving the jelly-like appear-
ance.

When the ferment is grown in solid gelatin and malt-wort,
spherical colonies are formed, but these appear only on the surface
or a short distance from it. The colonies that grow on the surface
gradually spread out into a film, resembling the ordinary membrane
which grows on nutrient liquids. Gelatin is not liquefied by the
ferment.

The membranous growth of the “vinegar plant” is the only
form of growth which I have succeeded in developing. During my
work with pure growths, I have made successfully more than 100
cultivations in solutions often varying very much in their chemical
composition ; but however much the character of the nutrient fluids,
or other conditions of growth, were altered, the ferment always
reproduced itself in the same very characteristic membranous form
described above.

In the many pure cultivations of B. aceti I have made, no form of
growth in the least resembling in microscopic appearance the tough
gelatinous film of the “vinegar plant” was ever observed; the
surface zoogloea form of B. aceti being thin and easily broken up by
the least agitation. The chemical reactions of the two growths are
also entirely different ; a cold solution of potash at once completely
disintegrates the pellicle of B. aceti, but the “vinegar plant”
membrane will withstand boiling with this solution for many hours

434 BROWN ON AN ACETIC FERMENT

without appreciable change. The pellicle of B. aceti on treatment
with concentrated sulphuric acid and iodine, gives no colour, but the
“vinegar plant” is stained of a deep blue, similar reactions to
these being also given by iodine dissolved in zinc chloride solution.
Thus the membrane of the “vinegar plant” gives the characteristic
reactions for cellulose (of which I shall show further on it is chiefly
composed), whereas the zoogloea pellicle of B. aceti does not show any
trace of this substance.

Previous writers who have described the membranous growth of
the “vinegar plant” (e.g., Zopf, Die Spaltpilze, p. 63, who evidently
refers to it under the name “ Essighautchen,” and “ Essigmutter’’),
have considered it to be a zooglaa form of B. aceti; but, from what
I have said above, it will be seen that this is open to the strongest
doubt. The true zoogloa form of B. aceti is an entirely different
form, as I have just shown. Considering, therefore, that under so
many different conditions of growth, the “vinegar plant” adheres to
its one distinctive form, without ever approaching to that of B. aceti,
and considering also that it develops a distinct chemical compound,
cellulose, which appears to be entirely absent from B. aceti, it is
impossible to come to any other conclusion than that the two
organisms are specifically distinct.

When a membrane of the “ vinegar plant” is examined microscopic-
ally, it is fuund to consist of bacteria, arranged more or less in lines, and
lying embedded in a transparent structureless film. These bacteria are
most commonly found as rods about 2 » in length, several often being
united together. The divisions, however, are readily seen when the
growth is dried and stained with aniline-violet, which brings out the
bacteria very distinctly, as the membrane in which they lie remains
colourless. In old cultivations, the rods are often to a large extent
replaced by micrococci about 0'5 « in diameter (? spores). Sometimes,
and more especially when the ferment has been grown in an unsuitable
nourishing medium, like yeast-water, it appears as long twisted threads
from 10—30 » in length, and of a leptothrix nature. I have never
observed with this ferment the peculiarly swollen involution forms so
frequent in old cultivations of B. aceti. In the upper surface of an old
membrane of the “ vinegar plant,” and also in the membrane grown
on the surface of solid gelatin, cells are often found whose sides are
much distended, and contain one or two highly refrangent nuclei
resembling spores in appearance. These nuclei (or spores) are deeply
stained by aniline dyes, and are often found free from any envelope.

In the liquid in which a membrane of ferment is growing, a few
free swimming cells are always to be found, particularly if the mem-
brane has been shaken, but these cells on cultivation always reproduce
the membranous growth.

WHICH FORMS CELLULOSE. 435

A temperature of about 28° appears to be most favourable to the
growth of the “vinegar plant.” Above 36°, it refuses to grow, but
still lives.

The fermentations produced by this ferment show the same chemi-
cal changes as those produced by B. aceti, so far as I have investi-
gated them. Ethylic alcohol is oxidised by it to acetic acid, and the
acid so formed is afterwards entirely broken up. Dextrose is oxidised
to gluconic acid; and levulose is also formed from mannitol. Like
B. aceti also, it has no fermentative action upon cane-sugar, starch, or
levulose. These results were arrived at in a similar manner to those
obtained with B. aceti, and fully described in my previous paper, it
is therefore needless to describe them further.

The great chemical difference between the “vinegar plant” and
B. aceti is, as we have seen above, the presence of a membrane hold-
ing together the cells of the ferment, and which gives the usual
reactions for cellulose. In order to ascertain with certainty the
chemical composition of this substance, I proceeded in the follow-
ing manner. A membrane of the ferment was taken, and after
well washing with hot water, was boiled for 20 minutes with a
10 per cent. solution of caustic potash. This treatment did not
appreciably affect the gelatinous film, but the bacteria embedded in
it were disintegrated. After washing the film with dilute hydro-
chloric acid, and afterwards with water, it was treated with a solution
of bromine, according to Miiller’s process for obtaining pure cellulose.
After following out this method, the final product was a colourless
semi-transparent film, retaining the shape and gelatinous character of
the original membrane. When examined under the microscope, no
trace of structural form could be detected. On treating a portion
with ammonio-cupric oxide solution, it dissolved with ease, and the
filtered solution, when acidified with hydrochloric acid, gave a pre-
cipitate exactly similar tu that from the cellulose of cotton-wool when
treated in a similar manner. Strong sulphuric acid dissolved the
membrane without blackening, and on diluting the solution with water
and boiling, a sugar was formed which reduced Fehling’s solution.

A portion of the membrane was dried at 105°, and 0°2302 gram
weighed out, and burnt with chromate of lead. This yielded 0°3737
gram CO, and 0°1295 gram UH:. On calculation this gives—

“cc

Cellulose
(CgH 095) ns
4444

6°17
49°39

100°00

436 BROWN ON AN ACETIC FERMENT

This analysis, together with the reactions mentioned above, leave
no doubt that the membrane of the “ vinegar plant” is cel/ulose. As
the cellulose in a fresh membrane of the ferment is easily dissolved
by ammonio-cupric oxide, this, according to Fremy and Urbain
(Compt. rend., 93, 926), shows it to be cellulose proper, like that of
cotton-wool; metacellulose, usually found in fungi, is insoluble under
similar circumstances, even after treatment with acids. The envelope
of yeast cells, usually said to be cellulose, is (according to Liebig)
also insoluble in ammonio-cupric oxide.

A different ferment from the one with which we are now concerned,
and known as Leuconostoc mesenterioides, is described as being
enveloped in a gelatinous membrane somewhat similar in appearance
to the membrane of the “vinegar plant;”’ but Scheibler has shown
that this body is an insoluble modification of the gum, dextran, and
decomposable by alkalis into the soluble form. I have examined
both the membrane of the “ vinegar plant,” and also the solation in
which it grows, but can find no trace of dextran.

Quantitative estimations (by Miiller’s process) of the cellulose in
membranes of the “vinegar plant,” grown under various conditions,
show that it varies from 35 to 62 per cent. when calculated on the
total weights of the original membranes dried at 100°.

This production of cellulose by a simple cell plant, and its use as a
cell connecting medium, seems of great interest in view of the impor-
tant part which cellulose plays in a similar manner in the more
highly organised forms of the vegetable kingdom; and it appeared
that any information that could be gained, as to the materials from
which cellulose is formed by the “viuegar plant,” might perhaps
assist in better understanding the complex reactions which go on in
the higher plants. To this end, my first experiments were made by
inoculating the “vinegar plant” into sterilised solutions of Pasteur’s
mineral medium, containing respectively 3 per cent. of cane-sugar,
dextrose, starch, and alcohol. These solutions were kept at a tempe-
rature of 28° for six weeks, but, with the exception of the solution
containing dextrose, no growth took place. The dextrose solution
developed a small but well-marked growth of the membrane. This
set of experiments gives fairly good proof that dextrose can be con-
verted into cellulose by the ferment; but the negative proof concern-
ing cane-sugar, starch, and alcohol, is weak, as some cause other
than the inability of these compounds to be converted into cellulose,
might have prevented the ferment from growing. In order to get
over this difliculty, recourse was had to yeast-water, in which solution
the ferment is able to grow, but only, as I have stated above, in a
very weak form. 1 gram of each of the carbohydrates to be used
(viz., cane-sugar, dextrose, and starch), was dissolved in 100 c.c. of

WHICH FORMS CELLULOSE. 437

the same preparation of yeast-water, sterilised, and inoculated with
the ferment; and at the same time two flasks, containing 100 c.c.
each of the yeast-water alone, were inoculated. After 16 days, the
flasks were opened and examined; the flask containing dextrose had
developed a thick white membrane on the surface, but all the others
showed a transparent jelly-like growth in the body of the liquid.
The films of ferment in all the experiments were then removed care-
fully, and after washing with water, were treated with dilute potash
and afterwards by Miiller’s process, in the manner described above.
The pure cellulose films thus obtained were then dried at 100°, and
weighed, with the following results :—

Weight of cellulose.
Dextrose and yeast-water 0°0227 gram.
Starch and yeast-water
Cane-sugar and yeast-water....
Yeast-water alone No. I
- » No.IL...... 00052

In considering these experiments, it is evident, from the extremely
close agreement of the weight of cellulose derived from the two
solutions of yeast-water alone, that 5°2 mgrms. may safely be taken

as the amount due to the yeast-water when the ferment had developed
in it. We can therefore fairly deduct this from the weights found in
the other experiments, and thus ascertain how far the carbohydrates
contained in the solutions may have been converted into cellulose.
Thus we find that 17°5 mgrms. of cellulose had been formed from
dextrose, whilst only 15 mgrms. of cellulose had been formed in
the starch solution, and 2°8 myrms. in the cane-sugar solution.
These latter quantities are so very small, that when we consider
that it is very difficult to obtain starch quite pure, and that
the cane-sugar used is sure to have been slightly inverted by the
repeated boilings necessary for sterilisation, I think we may safely
conclude that the “vinegar plant” is unable to convert either starch
or cane-sugar into cellulose, even when growing freely in their
presence. The weight of the cellulose formed in the dextrose solu-
tion quite confirms the first experiment with Pasteur’s solution, in
showing that the ferment can convert this sugar into cellulose. The
experiments I have just described are only one series out of two
which I made, both of which pointed to a similar conclusion.

The formation of cellulose from dextrose by the ferment is note-
worthy, as we have here a case of a simple cell having two totally
distinct actions on the same chemical compound ; for, as I have before
said, the fermentative action of the cell upon dextrose is to produce
gluconic acid by a fixing in some way of the oxygen of the air, whilst

438 BROWN ON AN ACETIC FERMENT

it also has the power of constructing the more complex molecule,
cellulose, from the same substance. It is of course impossible to
determine whether any one cell has these two powers at the same
time, but I have determined by experiment that during the time that
the cellulose membrane is growing, gluconic acid is formed.

The formation of cellulose by the “vinegar plant” from the
carbohydrates must not be looked on as an act of fermentation, as it is
evidently only a product of assimilation formed for the special use of
the plant itself.

A series of experiments carried on in a manner similar to the one 1
have just described, was made to ascertain the action of the ferment
on mannitol and on levulose, the results of which are shown in the
following table. 100 c.c. of the same yeast-water was used in each
experiment :— .

Total weight | Cellulose in | Per cent. of

of film, | film, | cellulose in
dried at 100°. | dried at 100°. | film.

Dextrose (CcHin0, + levees ,
2 grams.. ee eeeees

Mannitol (C;H,,0,), 2 grams..

Leevulose from inulin, 2 grams.

Levulose from mannitol, 2
GIONS ccccecccecccscccece

Yeast-water alone

It is evident, from the above experiments, that mannitol and
levulose are far more favourable to the growth of the ferment
than dextrose, and also that cellulose is formed from them more
freely.

During the growth of the ferment in levulose solutions, no trace of
fermentative action was observed (neither acid nor alcohol was
formed). In Experiment 4, an estimation of the levulose was made in
the solution with the polariscope, both before and after the growth
of the membrane. 9°314 gram of levulose had disappeared, whilst a
membrane of ferment weighing 0°1643 gram, and containing 0°1015
gram of cellulose, had grown in the liquid. Not less than 0°113 gram
of the sugar disappeared would be required for the formation of
the cellulose found, the remainder being probably used for the growth
of the ferment cells. The great difference between this experiment
with levulose and the experiment with dextrose—where true fer-

WHICH FORMS CELLULOSE. 439

mentation had been going on—will be evident by comparing the
results of the analyses of No. 1 experiment with the above.

Here 1:111 gram of dextrose was decomposed, 0°720 gram of which
was found as gluconic acid, the remainder having gone for the pro-
duction of cellulose, food for the cells, &¢. Yet—notwithstanding
active fermentation having been carried on—the total weight of the
ferment formed was only about one-fifth of that found when the
ferment grew in a solution of levulose, although no fermentation was
apparent in the latter experiment. This seems to raise a question as
to whether the “ vinegar plant” derives any benefit itself from the
oxidising fermentative power it possesses when grown in presence of
dextrose, alcohol, &c., and whether this property may not be some-
thing entirely unconnected with the necessities of its growth. My
experiments appear to show the possibility of such being the case, and
I intend to investigate these actions further.

During the growth of the ferment in presence of mannitol, this
substance is converted into levulose; it is therefore not possible to
say that mannitol itself can be converted directly into cellulose.

An experiment was made to ascertain if ethylic alcohol took part in
the growth of the cellulose membrane of the “vinegar plant.” In
this experiment the ferment was grown in yeast-water containing
4 per cent. of alcohol ; a cultivation being made in yeast-water alone
at the same time. At the end of the experiment, the same weight of
cellulose had been formed in both cases, thus showing that alcohol
took no part in its formation. In the alcoholic solution, 1-20 gram of
acetic acid was found.

The “vinegar plant” is frequently used in country places for the
manufacture of home-made vinegar, by introducing a membrane of
the ferment into a solution of coarse brown sugar. As a pure culti-
vation of this ferment has no action on cane-sugar, I procured a
membrane that had been used for the purpose described above to
examine its action. Microscopic examination showed the ferment to be
much contaminated with ordinary yeast cells (Saccharomyces cerevisie).
On introducing a portion into a cane-sugar solution, the yeast cells
present inverted and fermented the sugar, forming alcohol which,
after the first fermentation had ceased, was oxidised to acetic acid in
the ordinary way by the growth of a membrane of the vinegar plant
on the surface of the solution.

The “ vinegar plant” has no distinctive scientific name, I therefore
suggest, in consideration of its power of forming cellulose, that
Bacterium xylinum would be a suitable name for this ferment.

440

XLIV .—Puranitrobenzoylacetic Acid and some of its Derivatives.
By W. H. Perkin (Jun.), Ph.D., and Gustav Betienor, Ph.D.

In previous researches on ethylic benzoylacetate, it has been shown
that it is not only capable of entering into almost all the reactions in
which ethylic acetoacetate takes part, but moreover has the great
advantage that its derivatives, being aromatic substances, have a
much greater tendency to crystallise than those of the acetoacetate.

During the study of some of the more complicated derivatives of
ethylic benzoylacetate, and more especially the benzoyl-derivatives of
trimethylene and tetramethylene, it was found necessary, in order
to explain the formation of these compounds, that not only the acids
themselves, but also their ethereal salts should be solid, but this is
not always the case when ethylic benzoylacetate is employed.

It seemed, therefore, if a nitro-derivative of ethylic benzoylacetate
could be employed as the starting point instead of the ethyl salt
itself, that this difficulty could be avoided, and such is, in fact, the
case.

Experiments were first made with the object of nitrating ethylic
benzoylacetate directly, but as it is difficult to obtain the latter
free from acetophenone, benzoic acid, and other impurities, no satis-
factory results could be obtained, and this mode of procedure was
abandoned.

It was, however, soon discovered that paranitrobenzoylacetic acid
could easily be prepared by a process analogous to that for the forma-
tion of benzoylacetic acid from phenylpropiolic acid; that is, by
treating paranitrophenylpropiolic acid with concentrated sulphuric
acid :—

NO,C,H,’C :C-COOH + H,O = NO,C,H,CO-CH,-COOH

Paranitrophenylpropiolic acid. Paranitrobenzoylacetic acid.

This fact is all the more interesting when it is remembered that
ethylic orthonitrophenylpropiolate behaves so differently when
treated in the same way. Baeyer (Ber., 14, 174), who studied this
reaction, has shown that on treating the ortho-derivative with sul-
phuric acid, a simple intramolecular change takes place, ethylic

isatogenate being formed :—
O,

ii.
,, os COOC,H;.

Ai<vo CoH | /

Ethylic ort aiid, Ethylic isatogenate.
propiolate.

0: C-0000H.

a. 2 - & 2 22 Oe Ge ae ee OO

PERKIN AND BELLENOT: PARANITROBENZOYLACETIC ACID. 441

It seemed of interest, therefore, to prepare some of the derivatives
of paranitrobenzoylacetic acid, an account of which we now desire to
lay before the Society.

We have divided this research into the following sections :—

(1.) Paranitrophenylpropiolic acid.

(2.) Paranitrobenzoylacetic acid.

(3.) Methylic paranitrobenzoylacetate.

(4.) Methylic benzylparanitrobenzoylacetate.
(5.) Ethylic paranitrobenzoylacetate.

(6.) Ethylic nitrosoparanitrobenzoylacetate.
(7.) Ethylic ethylparanitrobenzoylacetate.
(8.) Ethylic allylparanitrobenzoylacetate.
(9.) Ethylic di-paranitrobenzoylsuccinate.

(1.) Paranitrophenylpropiolic Acid.

Paranitrocinnamic acid and its derivatives were first carefully
examined by Drewsen (Annalen, 212, 150).

If this acid is treated with bromine, it easily takes up 1 mol. Br
exactly in the same way as cinnamic acid does, and is converted into
paranitrophenyldibromopropionic acid, NO,-C,H,CHBrCHBr-COOH.

If this dibromo-acid is treated with alkalis, a curious reaction takes
place, no a- or f-monobromo-acid being formed as in the case of
dibromophenylpropionic acid, but the dibromide is split up partly
into paranitrophenylpropiolic acid and partly into the original para-
nitrocinnamic acid, bromine being given off. In preparing para-
nitrophenylpropiolic acid by this method, the separation of these two
acids is not only a matter of considerable difficulty, but the yield of
acid obtained is very small. It was found, however, that better results
could be obtained by using the dibromide of the ethereal salt instead
of the free acid.

The following method of preparation is based on that proposed by
Drewsen :—

Ethylic paranitrophenyldibromopropionate is first prepared by
mixing pure ethylic paranitrocinnamate with a slight excess of bro-
mine. On pounding the substances together, the ethereal salt dis-
solves in the bromine, and in a short time the whole solidifies to a
brown cake of impure ethylic paranitrophenyldibromopropionate.
This is reduced to powder, spread out in the air till the excess of
bromine has evaporated, and then crystallised once from alcohol.
In this way it is obtained quite pure in colourless crystals, and
can be used directly for the preparation of paranitrophenylpropiolic
acid.

For this purpose, the crystals are dissolved in a small quantity of

VOL, XLIX. 2H

442 PERKIN AND BELLENOT: PARANITROBENZOYLACETIC ACID

boiling alcohol, and the calculated amount of a hot alcoholic potash
solution poured in until the mixture has a decidedly alkaline reaction.
As soon as the reaction is finished, a slight excess of potash causes
the solution to become brownish coloured, probably from the forma-
tion of azo-derivatives. After remaining for an hour or so, about
twice the volume of water is added, the whole filtered, and the para-
nitrophenylpropiolic acid precipitated by acidifying with dilute sul-
phuric acid.

The precipitate being very difficult to collect, it is better to agitate
the whole repeatedly with ether. By this means also a purer product
is obtained, as the paranitrocinnamic acid which is always formed
during the reaction is not dissolved by the ether. The ethereal solu-
tion is well washed with water and shaken with dilute potash
solution, when the paranitrophenylpropiolic acid dissolves, leaving
the paranitrophenylacetylene and paranitroacetophenone—which are
always present—in the ethereal solution. The acid is then repre-
cipitated from the aqueous potash solution and again extracted with
ether.

On distilling off the ether, the paranitrophenylpropiolic acid is
obtained pure in small needles, and can be used directly for the
preparation of paranitrobenzoylacetic acid.

(2.) Paranitrobenzoylacetic acid.

In order to prepare ethylic paranitrobenzoylacetate, experiments
were first made on the action of sulphuric acid on ethylic paranitro-
phenylpropiolate. 50 grams of the pure ethereal salt were dissolved
in 2 kilos. of concentrated sulphuric acid which had been diluted with
about 40 c.c. of water, and the whole allowed to remain for several
hours at the ordinary temperature. As no reaction appeared to take
place at this temperature, experiments were made with a view of
ascertaining the proper conditions for the transformation, and it was
ultimately found that the addition of water took place best at 35—40°,
and required at least 12 to 15 hours before the reaction was completed.
On pouring the product on to ice, adirty whitish flocculent precipitate
was formed, which was collected, well washed with water, and dried
ona porous plate. The addition of a drop of ferric chloride to the
alcoholic solution of this substance gave a beautiful violet coloration,
a proof that the reaction had really taken place.

It was soon discovered, however, that the compound formed was
not the ethereal salt of paranitrobenzoylacetic acid, but the acid itself,
the ethereal salt having undergone hydrolysis by the action of the
sulphuric acid. All attempts to obtain the ethereal salt directly were
in vain.

AND SOME OF ITS DERIVATIVES. 443

As the yield of acid obtained in this way was not good, it was
necessary in order to be able to prepare larger quantities to improve
the method. After numerous experiments, the following—which we
give in detail—was found to give the best results.

Pure paranitrophenylpropiolic acid (50 grams) is dissolved in 2—3
kilos. of 80—85 per cent. sulphuric acid, and the whole kept at a
temperature of 35—40° for 15 hours. The solution which is at first
yellowish-coloured becomes darker and darker, and is at last of a deep
brown. In order to determine whether the reaction is finished a
sample is taken out, mixed with ice and water, and extracted with
ether. The residue left on distilling off the ether is dissolved in
dilute ammonia and boiled for about two minutes; this causes the
separation of oily drops of paranitroacetophenone, which on cooling
become solid and can be filtered off. On acidifying the filtrate, any
unchanged paranitrophenylpropiolic acid is precipitated, and from
the amount of this it is easy to judge how far the reaction has pro-
gressed.

When it is found that the reaction is finished, the product is poured
on to ice. This causes the paranitrobenzoylacetic acid to be precipi-
tated in whitish flakes, which must be removed from the acid liquid
as rapidly as possible by extraction with ether. The ethereal solu-
tion is well washed with water, dried over calcic chloride, and the
ether distilled off until crystals begin to form; these consist of
nearly pure paranitrobenzoylacetic acid. The filtrate from these
crystals contains, besides this acid, unchanged paranitrophenylpro-
piolic acid and paranitroacetophenone; these can be separated by
recrystallisation from benzene. Pure paranitrobenzolyacetic acid
crystallises from this solvent in almost colourless microscopic needles.

Analysis :—
I. 0°2511 gram substance gave 0°0840 gram H,0 and 0°4698 gram
CO,.
II. 0°2292 gram substance gave 13°2 c.c. N at 16° and 720 mm.
IIf. 03165 ,, . » 19°0 cc. N at 13° and 709 mm.

Found.
scat dilinesi tained Theory.

I. 10 NO,-C,H,-CO-CH-COOH.
G.n00 ; — — per cent. 51°19 per cent.
H... 37: mn 334,

”

Bice 635 6°61 - 6°69 -

This acid is formed from paranitrophenylpropiolic acid according

to the equation—

NO.-C.HyC : C-COOH + H,O = NO,C,H,-CO-CH-COOH.

444 PERKIN AND BELLENOT: PARANITROBENZOYLACETIC ACID

Paranitrobenzoylacetic acid melts at 135°, being at the same time
decomposed into paranitroacetophenone and carbonic anhydride. It
is easily soluble in alcohol, ether, bisulphide of carbon, and light
petroleum, more sparingly in benzene. The alcoholic solution gives an
intense reddish-brown coloration with ferric chloride.

If the finely powdered acid is warmed with water at 60°, it dissolves
in considerable quantity and without much decomposition, crystallis-
ing out on cooling in four-sided plates. The hot aqueous solution is
coloured deep reddish-brown on the addition of ferric chloride, the
cold saturated solution, however, only very slightly so.

When boiled with water or dilute acids, the acid is rapidly decom-
posed into carbonic anhydride and paranitroacetophenone :—

NO,.C,H,CO-CH,,COOH = NO,C,HyCO-CH; + CO.

The preparation of pure salts of paranitrobenzoylacetic acid is
a matter of great difficulty owing to their instability. The acid
dissolves easily in dilute alkalis, with an intense yellow colour, the
solution, however, soon becomes filled with needles of paranitro-
acetophenone.

The solution of the sodic salt was prepared by carefully neutralis-
ing the solution of the acid in sodic carbonate with dilute nitric acid.
On the addition of silver nitrate to this, a light yellow amorphous
silver salt was obtained, which was very unstable, however, and on
analysis gave approximate numbers only.

The copper salt is a light green amorphous salt, which explodes if
heated in the dry state.

The other salts are also very unstable.

(3.) Methylic Paranitrobenzoylacetate.

The methylic salt of paranitrobenzoylacetic acid is easily obtained
by suspending the pure acid in absolute methyl alcohol, and saturat-
ing with hydrochloric acid gas, the whole being well cooled during the
operation. In a short time, the methylic salt begins to crystallise out
in yellow needles. After remaining for 12 hours at as low a tempera-
ture as possible, the crystals are collected and well washed with methyl
alcohol. By one recrystallisation from methyl alcohol, they are easily
obtained pure.

Analysis :—

02360 gram substance gave 0:0926 gram H,0 and 0°4646 gram CO:.

Theory.
NO,-C,H, -CO-CH,-COOCH;.

53°81
4°04

AND SOME OF ITS DERIVATIVES. 445

Methylic paranitrobenzoylacetate melts at 106—107°. It is only
sparingly soluble in cold ethylic or methylic alcohol, but easily on
boiling. It is easily soluble in light petroleum, benzene, and chloroform.
The addition of a drop of ferric chloride to the alcoholic solution gives
a brownish-violet coloration.

If the hot alcoholic solution of this ethereal salt be allowed to cool
slowly, beautiful well-formed crystals can be obtained. For the
following description of these I am indebted to the kindness of Pro-
fessor Haushofer.

Methylic Paranitrobenzoylacetate.

Crystalline Form. Monoclinic.
a:b:c = 0°4993: 1 (?).
B = 56° 5’.
The crystals form the prismatic twinned combination 0P(c), coP( p)
in the direction of the vertical axis.

Measured.
p:p = *135° 0’
¢:p = *121° 3’

On the plane p, the angle made by the directions of extinction (with
crossed Nicols) with the edge of the prism is about 15°.

Sodium compound, NO.°C;H,yCO-CHNa:COOCH;.—If the solution of
methylic paranitrobenzoylacetate (1 mol.) in warm alcohol is treated
with the calculated quantity (1 mol.) of a strong solution of sodic
ethylate or methylate, a dark orange-brown solution is formed, and,
on stirring, the whole solidifies to a mass of yellow crystals. These
after collecting, well washing with alcohol, and drying on a porous
plate in a vacuum, gave the following numbers on analysis :—

0°2690 gram substance gave 0°0768 gram Na,SQ,.

Theory.
Found. NO,'C,H,-CO-CHNa-COOCH;,.
9°27 per cent. 9°38 per cent.
2H 2

446 PERKIN AND BELLENOT: PARANITROBENZOYLACETIC ACID

The substance is, therefore, the monosodium-derivative of methylic
paranitrobenzoylacetate. It is extremely stable, and behaves more
like the salt of an acid than a sodium-derivative. It dissolves in
water without decomposition, forming a yellow solution. It is
scarcely attacked when treated with iodine.

(4.) Methylic Benzylparanitrobenzoylacetate.

It seemed to be of interest to determine whether the paranitro-
derivatives of benzoylacetic acid could be used for synthetical experi-
ments with the same advantage as benzoylacetic acid itself.

For this purpose, the action of benzyl chloride on the sodium-deri-
vative of methylic paranitrobenzoylacetate was studied. The pure
sodium-derivative was heated with the calculated quantity of benzyl
chloride and a little alcohol in a sealed tube for six hours at 150°.
The product was mixed with water, and once or twice extracted with
ether. After drying the ethereal solution and distilling off the ether,
a thick brownish oil remained, which did not solidify even when
exposed for some days over sulphuric acid in a vacuum.

On distilling this crude product in a current of steam, a small
quantity of oil passed over, but the main part remained behind in the
retort. This was extracted with ether, and the ethereal solution well
dried; on distilling off the ether, a thick oil remained, which, how-
ever, soon solidified to a cake of crystals. These were roughly purified
by spreading them out on a porous plate, and then once or twice re-
crystallising from dilute alcohol. Analysis :—

I. 0°1728 gram substance gave 0°4113 gram CO, and 0°0755 gram
H,0.
II. 0°3150 gram substance gave l3c.c.N. ¢= 11°. Bar. = 710 mm.

Found.

Theory.
II. NO,-C,H,-CO-CH (C;H,)-COOCH,.
— per cent. 65°17 per cent.
Kn ” 4°79 ”
4°57 - 4°47 -

The substance was therefore methylic benzylparanitrobenzoylacetate,
formed according the equation—

NO,-C,H,-CO-CHNa-COOCH, + C;H,Cl =
NO,-C,H,CO-CH(C,H;)-COOCH, + NaCl.

It dissolves easily in most solvents, and crystallises from dilute alcohol
in colourless plates which melt at 57°.

AND SOME OF ITS DERIVATIVES.

(5.) Ethylic Paranitrobenzoylacetate.

In order to prepare this ethereal salt, pure paranitrobenzoylacetic
acid is suspended in an excess of absolute ethyl alcohol and the liquid
thoroughly saturated with hydrogen chloride, care being taken that
the temperature does not rise above 20° during the operation, other-
wise decomposition is apt to set in. The product, which is a clear
brownish liquid, is allowed to remain for two hours, and then poured
into ice and water. This precipitates the ethereal salt in oily drops
which solidify after a time; these can be collected on a filter, but it
is best to extract the whole with ether. The ethereal solution, after
washing with water, drying over calcic chloride, and distilling off the
ether, deposits a thick brownish oil, which gradually solidifies on
standing over sulphuric acid in a vacuum; this crude product can
easily be separated from paranitroacetophenone and tarry products, by
dissolving it in a little boiling alcohol.

On cooling, the ethylic paranitrobenzoylacetate crystallises out in
yellow needles, which may be obtained almost colourless by repeated
recrystallisation. The analysis gave the following results :—

I. 0°1619 gram substance gave 0°0672 gram H,0 and 0°3304 gram
CO,.
II. 0°1993 gram substance gave 10°50 cc. N. ¢=12°. Bar.=

709 mm.

Found.

Theory.
II. NO,-C,H,-CO-CH,-COOC,H,.
— per cent. 55°64 per cent.
nine ” 4°64 ”

5°83 ” 5°90 ”

Ethylic paranitrobenzoylacetate is easily soluble in boiling alcohol,
much less so in the cold; it crystallises, however, best from a mixture
of benzene (1 part), and light petroleum (b. p. 6(0—80°, 2 parts). If this
solution be allowed to evaporate at the ordinary temperature, beautiful
transparent crystals are obtained, melting at 74—76°. Professor
Haushofer, who was kind enough to measure these, gave me the follow-
ing account of them.

Crystalline Form. Monoclinic.
a:b:¢ = 0°3584: 1: 1:2380.
8 = 72° 22’.
Tabular prismatic crystals developed in the direction of the
clinodiagonal, with the combinations 0P(001) = ¢, Roo(011) = q,
coPco(100) = a, — Poo(101) = 7, — 2P2(121) = ¢, — 4P4(141) = s.

448 PERKIN AND BELLENOT: PARANITROBENZOYLACETIO ACID

The surface c well developed but usually rough; the surfaces ¢ and
s usually very small, scarcely visible to the naked eye.

Measured. Calculated.
ce: a = (001)(100) = *107° 38’ _ —
q:q = (011)(011)= *80 34 — — (above)
e:7r = (001)(101) = *121 52 _ —
t:r = (121)(101)= 148 10 148° 40’
s:r = (141)(101)= 128 33 129 24
gq: a = (011)(100)= 101 26 101 28

The plane of the optical axes issymmetrical. In convergent polarised
light, the interference image of an axis makes its appearance on the
surface c.

Ethylic paranitrobenzoylacetate dissolves fairly easily in sodic
hydrate, especially on warming, pfoducing a yellow solution. On the
addition of acids, it is reprecipitated unchanged. Its alcoholic solution
gives a beautiful brownish-violet coloration with ferric chloride.

Ethylic paranitrobenzoylacetate is easily reduced when treated with
tin and hydrochloric acid in aleoholic solution, but the products formed
have not been further examined.

If a strong solution of sodic ethylate (1 mol.) be added to the
alcoholic solution of ethylie paranitrobenzoylacetate (1 mol.), an
intense brownish-yellow solution is formed, which on standing solidifies
to a mass of orange-coloured crystals of the sodium compound. These
when collected, well washed with alcohol, and dried over sulphuric
acid in a vacuum, gave the following numbers on analysis :—

0°2583 gram substance gave 0°0705 gram Na,SQ,.

Theory.
Found. NO,-C,H,:CO-CHNa-COOC,H;.
MGs ccces 8°86 per cent. 8°88 per cent.

The dry sodium compound is characterised by its intensely bitter
taste. It explodes feebly when heated, paranitroacetophenone and
other compounds being formed.

AND SOME OF ITS DERIVATIVES. 449

It is sparingly soluble in cold water, but easily on warming, form-
ing a deep yellow solution, which on the addition of an acid is in-
stautly decolorised, ethylic paranitrobenzoylacetate being precipitated.
If the solution of the sodium compound is mixed with a solution of an
inorganic salt, the former is completely precipitated and the solution
becomes colourless. The sodium compound is fairly easily soluble in
boiling alcohol, and crystallises out on cooling in small orange-yellow
needles, which are much redder than the sodium compound of methylic
paranitrobenzoylacetate. The addition of ferric chloride to the alco-
holic solution produces a dark reddish-brown coloration; but if this
reagent is added to the solution of the sodium compound in water, a
light reddish precipitate is formed, which appears to be crystalline, and
is probably the ferric salt of ethylic paranitrobenzoylacetate.

If nitrate of silver is added to the solution of the sodium compound
in water, a light yellow, amorphous precipitate is formed which ex-
plodes when heated in the dry state. The lead compound is easily
obtained as a yellow amorphous mass, on adding a solution of acetate
of lead to the aqueous solution of the sodium compound. The copper
compound is a yellowish-green precipitate.

The compounds of ethylic paranitrobenzoylacetate with metals are
remarkable for their stability. The sodium compound behaves
exactly like the salt of an acid, it dissolves in water without decom-
position and can even be recrystallised from it. The sodium compound
of ethylic benzoylacetate is decomposed by boiling with water, and
ethylic sodacetoacetate is split up when merely mixed with water.
The sodium compound of methylic paranitrobenzoylacetate is, curiously
enough, still more stable than that of the ethylic salt.

(6.) Ethylic Nitrosoparanitrobenzoylacetate.

As has been shown in previous researches (Trans., 1885, 244), ethylic

benzoylacetate is easily converted into its isonitroso-derivative,
C,H;-CO-C(NOH)-COOC.H;,

when its solution in dilute sodic hydrate is treated with sodic nitrite

and dilute sulphuric acid. This method, however, is not applicable

to ethylic paranitrobenzoylacetate.

If the solution of the sodium compound of the ethereal salt is
mixed with sodic nitrite, the former is precipitated almost entirely,
and thus escapes the action of the nitrous acid when the mixture is
acidified with dilute sulphuric acid.

After several experiments, however, it was found that the desired
isonitroso-derivative could easily be obtained iu the following way :—
Pure ethylic paranitrobenzoylacetate is dissolved in dry ether and a

450 PERKIN AND BELLENOT: PARANITROBENZOYLACETIC ACID

stream of nitrous gas passed through the solution until it is saturated,
care being taken to cool well during the operation.

After standing for a short time, a rapid current of carbonic anhy-
dride is passed through in order to remove the excess of nitrous acid
as completely as possible, and the solution is allowed to evaporate
spontaneously. The semi-solid mass which remains is first roughly
separated from oily impurities by spreading it out on a porous plate,
and then twice recrystallised from dilute alcohol.

Analysis :—

I. 0°1646 gram substance gave 0°0504 gram H,O and 0°2980 gram
CO..
II. 0°1927 gram substance gave 19c.c.N. ¢= 14°. Bar. = 716mm.
Found.
-——s—- Theory.
I. IL. NO,-C,H,-CO-C(NOH)-COOC.H;.
49°37 — per cent. 49°62 per cent.
H... 340 ae 375,
NN... — 10°55, 10°53 -

This compound is, therefore, ethylic nitrosopararitrobenzoylacetate,
formed according to the equation—

NO.°C,H,CO-CH,-COOC,H, + NO.H =
NO,C,;H,CO-C(NOH)-COOC.H, + H,0.

It is easily soluble in alcohol, ether, and acetone, less readily in
benzene and light petroleum. It crystallises from dilute alcohol in
colourless needles, which melt at 220° with decomposition. It dis-
solves easily in sodic hydrate, forming a deep yellow solution, and is
reprecipitated unchanged on the addition of dilute acids.

(7.) Ethylic Ethylparanitrobenzoylacetate.

In preparing homologues of ethylic paranitrobenzoylacetate it is
always better to start from the pure sodium compound than from the
ethereal salt itself, as the products obtained in this way are always
purer.

In order to prepare ethylic ethylparanitrobenzoylacetate, the pure
sodium compound was heated with an excess of ethyl iodide and a
little alcohol for about three hours at 100°. The whole was then
evaporated to dryness on a water-bath, and the residue treated with
water and extracted with ether. On distilling off the ether, a thick
yellowish oil was left, which solidified after standing for some time
over sulphuric acid in a vacuum. The crystals were first roughly
freed from impurities by spreading the mass out on a porous plate, and
then recrystallised twice from dilute alcohol.

AND SOME OF ITS DERIVATIVES.
Analysis :—-

I. 01100 gram substance gave 0°0564 gram H,0O and 0°2365 gram
CO.
II. 0°1512 gram substance gave 8 c.c.N. ¢=13°. Bar. = 711mm.

Found.
— Theory.
I. Il. NO,:C,H,-CO-CH(C,H;)-COOC,H;.
58°63 — per cent. 58°86 per cent.
5°69 ees ” 5°66 ”
_ 5°50 7 5°28 99

Ethylic ethyl paranitrobenzoylacetate crystallises from dilute
alcohol in colourless plates which melt at 39—40°. It is easily
soluble in alcohol, ether, light petroleum, benzene, and carbon
bisulphide.

All attempts to obtain the corresponding ethyl paranitrobenzoyl-
acetic acid by hydrolysis were fruitless, paranitrobenzoic acid always
being formed, even when very dilute potash was used. No better
results were obtained with sulphuric acid at ordinary temperatures.

If ethylic ethylparanitrobenzoylacetate is heated with concentrated
sulphuric acid at 100°, carbonic anhydride is given off, and, on adding
water and extracting with ether, a yellowish oil is obtained; this on
long standing becomes solid. It appears to be ethyl paranitroaceto-
phenone, NO,°C,H,CO-CH,-C,H;.

The quantity was unfortunately too small for analysis.

(8.) Ethyl Allylparanitrobenzoylacetate.

This ethereal salt was prepared in exactly the same way as the
ethylic ethylparanitrobenzoylacetate, allyl iodide being used instead
of ethyl iodide. The crude product obtained was purified by repeated
crystallisation from dilute alcohol.

Analysis :—

I, = gram substance gave 0°1129 gram H,0 and 0°5100 gram

II. 0°1695 gram substance gave 79 cc. N. ¢= 14°. Bar. =

717 mm.

Found.
—~ Theory.
I. . NO,-O,;H,-CO-CH(CH,-CH:CH,)} ‘COOC;H;.
60°72 per cent. 60°64 per cent.
5°47 —_— * 41
_ 5°08 ” 5°05

”

Ethylic allylparanitrobenzoylacetate melts at 45—46°. A number

452 PERKIN AND BELLENOT: PARANITROBENZOYLACETIC ACID

of experiments were made with the object of obtaining the cor-
responding allylparanitrobenzoylacetic acid, but without result;
paranitrobenzoic acid always being formed. When heated with con-
centrated sulphuric acid at 100°, ethylic allylparanitrobenzoylacetate
is decomposed into carbonic anhydride and a ketone, probably allyl-
paranitroacetophenone, NO,°C,H,CO-CH,CH,CH : CH;.

(9.) Ethylic Diparanitrobenzoylsuccinate,
NO,.°C,H,-CO-CH:COOC.H,

| °
NO,-C,H,CO-CH-COOC,H;

This compound is formed when the sodium-derivative of ethylic
paranitrobenzoylacetate is treated with iodine:

2(NO,-C,H,CO-CHNa-COOC,H;) + I, =
NO.°C,H,CO-CH-COOC.H;

| + 2Nal.
NO,.-C,HyCO-CH:COOC,.H;

in a way exactly analogous to the formation of ethylic dibenzoylsuc-
cinate from the sodium compound of ethylic benzoylacetate. To
prepare it, the pure sodium compound of ethylic paranitrobenzoyl-
acetate is first finely powdered and then agitated with an ethereal
solution of iodine, containing the calculated quantity. If this solu-
tion be added by degrees, it will be seen that the colour of the iodine
disappears rapidly at first, but at last only after long shaking, and it
will be found usually that the whole amount is not decolorised.
The product is then mixed with water and sulphurous acid, the
ethereal solution separated, dried over calcic chloride, and evaporated.
In this way, a yellow oil is obtained which, after standing over
sulphuric acid in vacuum, deposits crystals. These must be
separated from the oily mother-liquor by spreading out on a porous
plate, and then recrystallised once or twice from dilute alcohol.
The analysis gave the following result :—

I. 0°2344 gram substance gave 0'0941 gram H,O and 0°4805 gram
CO,.

II. 0°2001 gram substance gave 10°77 cc. N. ¢=17°. Bar. =

712 mm.

Found.
—-A——, Theory.

VA 4 (NO,-C,H, -CO).C, Hy (COOC,H;)2-

C.... 55°90 — per cent, 55°93 per cent.

H... 448 — e 4°24 .

Bove

— 5°82 502,

”?

RAMSAY AND YOUNG: VAPOUR-PRESSURES OF BROMINE. 453

Ethylic diparanitrobenzoylsuccinate crystallises from dilute alcohol
in colourless needles which melt at 180°. It is easily soluble in
alcohol, ether, benzene, and light petroleum. If a solution of this
substance in dry ether is mixed with an ethereal solution of sodic
ethylate, a white amorphous compound is thrown down, which hes
without doubt the formula—

NO,C,H,yC O-CNa‘COOC.H;

| :
NO,.°C,H,-CO-CNa-COOC.H,

When ethylic diparanitrobenzoylsuccinate is treated with concen-
trated sulphuric acid, it is converted into a new compound, which is
probably ethylic diparanitrophenylfurfurandicarboxylate. Owing to
want of material, however, the reaction was not followed up.

From these experiments, it can be seen that ethylic paranitro-
benzoylacetate may be used with advantage instead of ethylic
benzoylacetate, the derivatives formed being solid and therefore easy
to purify.

In a short time, we hope to be able to communicate to the Society
the experiments on the formation of trimethylene- and _ tetra-
methylene-derivatives, which were made with this substance.

XLV.—On the Vapour-pressures of Bromine and Iodine, and on Iodine
Monochloride.

By Wittiam Ramsay, Ph.D., and Sypney Youne, D.Sc.

So far as we know, the vapour-pressures of bromine and iodine have
not been measured. The action of these elements on mercary precludes
the ordinary method of experiment; but it is possible to use the
method described by us in the Transactions, 47, 42, substituting for
the cotton-wool on the thermometer bulb a similar covering of
asbestos, and interposing, between the still-apparatus and the mano-
meter, a tube filled with solid sodinm amalgam.

Although the vapour-pressures of these elements have not been
determined, there are numerous isolated determinations of melting and
boiling points; these, however, differ so widely from each other that a
revision would seem desirable.

(1.) Melting Point of Bromine,—The following determinations have
previously been made :—

VOL, XLIX, 21

454 RAMSAY AND YOUNG: VAPOUR-PRESSURES

Regnault — 7°32° Balard — 18:0° (freezing point).
Pierre — 75 Serullas — 18° to 20°
Philipp) — 72 to — 73 Liebig — 250°

Quincke — 20°0

Baumhauer — 24°5°

Besides the statements of these authors, those of text-books are very
contradictory: for example, Roscoe and Schorlemmer, —21°; Miller,
—12°5°; Frankland, “ Lecture Notes,” — 20°; Wurtz’s Dictionary,
freezing point —22°, melting point above —12°.

(2.) Boiling Point of Bromine.—

Pierre, 63°0° at 760°32. Thorpe, 59°27° at 760.
Andrews, 58°0° at 760°0. Bolas and Groves, 59°6” at 751°0.
Landolt, 58°6° at ?

Besides these, Balard gives 47° and Liwig 45°, and one of these
numbers is quoted in Williamson (47°).
(3.) Melting Point of Iodine.—

Regnault, 107°. During solidification, 113°6°.
Stas, 1183—115°.

(4.) Boiling Point of Iodine.—
Regnault, 175° (about). Stas, above :200°.

Two sets of experiments were carried owt with samples of bromine
obtained from different sources. The first portion, after drying over
sulphuric acid and boiling with potassium bromide to remove chlorine,
distilled with absolute constancy from beginning to end at 57°65°* at
749°38 mm. The second sample was boiled with potassium bromide,
dehydrated with phosphoric anhydride, and boiled, after a small
quantity had come ever, constantly at 58°85°, at a pressure of
755°38 mm. Both of these samples were tested for iodine with nega-
tive results; and on adding 0°25 per cent. of iodine to a portion of one
of the samples, the boiling point was no longer constant. We are
unable to account for this difference in boiling point in the two
samples, which was found to be persistent throughout the whole
range of pressures.

The results with the first sample are as follows:—(b. p. 57°65° at
749°8 mm.).

* All the temperatures were determined with a mercurial thermometer, but are
corrected to express air-thermometer temperatures, for the mercuriul thermometers
had previously been carefully compared.

OF BROMINE AND IODINE.

Series I.

Temperature.

Pressure. Pressure.

Liquid. Liquid.

15 °98° 147°4 mm.
19 °30 174 °35

19-92 17545

28 -23 255°
30°25 275°
32°96 305
85°68 340°
38°57 379°
41 °47 421°
44°24 465°
46 *86 511
49 -22 558

lid |

NO AT

SLERSRSA °

~~ ee pe
ADONwSown
ot on

WOOK DORA

Temperature.

Pressure.

Solid. Liquid.

12°65°

12°27

11°02
9°9

Repeated with sample 2 (b. p. 55°85 at 755°8 mm.).

Temperature. Pressure. Temperature. Pressure.
Liquid.. — 9°88° 38°55 mm. — 2°63° 579 10m.
Solid .. — 9°68 35°55, — O31 65°25,
» «+ — 865 40°05 ,, + 10°4 1118 _—SC#,,,
Liquid... — 6:90 46°35, 28°55 2510 _,,
» «+ = 504 509 ,, 47°65 5182 __,,

With sample 1, series I, after the bromine had been frozen, on
admitting air to raise the pressure, the temperature remained stationary
at —7-05°; in series II, on solidification, the temperature rose from
—9°45° to —7-45° and then fell again to —9-9°; and when the
pressure was made to rise, the temperature again became stationary

212

456 RAMSAY AND YOUNG: VAPOUR-PRESSURES

at —7-0°. With sample 2, the temperature rose to —7°25° and then
fell again. From direct observation, therefore, the melting point is
seen to be between —7°0° and —7°45°.

As a comparison of these samples did not give concordant results,
the second sample was first treated with caustic soda, evaporated,
and the resulting bromate decomposed by heat. The bromine was
recovered by distillation with potassium dichromate and dilute sul-
phuric acid. By this treatment all organic impurity must have been
destroyed. The bromine was then digested with potassium bromide,
and finally dried with sulphuric acid. It boiled constantly at 58°65°
under a pressure of 757 mm.

A third sample, from a different source, was distilled alone, and
boiled at 56°8—58°5° at 757 mm. It was then digested at the boiling
point with potassium bromide, and allowed to stand over night; the
boiling point was then 57°8° to 58°45° at 750°9 mm. It was then
shaken with strong sulphuric acid and distilled at once, when it boiled
from 58°35° to 58°5° at 750°9 mm.; and on standing for two days with
strong sulphuric acid with occasional shaking, it boiled at 58°5° to
58°6° at 754°4 mm. This determination agrees with the result given
by the purified sample, hence the last sample was used for experiment,
and its vapour-pressures were measured. The rate of change of
temperature at the boiling point of bromine is very nearly 1° for
25 mm.; corrected to 760 mm., the various determinations are as
follows :—

First sample
Second sample

”

Third

The vapour-pressures of sample 3 were then determined.

Series I.

Temp. Pressure. Pressure.

53°8 mm. ' 80 °35 mm.
57°8 . 87°8
62°2 ' 97°65
65°0 ,, . 108 °8
73°95 ' 127 ‘25

org
wnesas
CAM TIS

%
§ g
F . ° ¢ ° . ° ° % re S
| &

750
+
750

700
nr

Journ. Chem.Soc. Sly I886
700

600

550
t+ +
at |
|
550 600

500

2 .
3 \ 3 8 i
\ “ & F
3
$ :

5
400
im

}
| cll]
wz
|
|
|
|
|
—*
—r
350
Pressure

S | )
S 7" i
3
= —f—_—_-4+ — —+-__-_-@ — ——} EE — —
S r -
S
o 3

200
}
T

400
=

——
+
|
|

CECE
7 : !
~ 2

AUTUDAST «= JO Saemypsaduiay

— >
os

T° +}--_-—+—

OF BROMINE AND IODINE. 457

Series IT.

Temp. | Pressure. Temp. Pressure. Temp. Pressure.
— 17°12° | 18°9 mm.]| — 0°85° 63°2 mm. 28 80° 250°55mm.
— 12°74 | 281 ,, + 2°78 75°75 ,, 32°70 293°4 ,,
— 12°70 | 27°65 ,, 6°38 ie 34°68 321°0_ ,,
— 10°36 | 34°85 _,, 14°30 133 “65 ,, 36°98 | 350°25 ,,
— 6°39 47°05 ,, 20°96 181 °25 ,, 38°61 375°75 ,,
— 4°79 me 22 -20 189 °75 ,, _ _

— 2°02 | 59°0_ ,, 25°72 220 °25 ,, _— _
Series IIT (with a Different Thermometer).
Temp. Pressure. Temp. Pressure. Temp. Pressure.

34°4° 315 ‘05 mm. 45°50° 478 ‘2 mm. 54°1° 636°1 mm.
37 *44 357°0 ,, 48 °7 540°5 ,, 56°0 689°0 ,,
41°85 418°6 ,, 50°2 567°3 ,, — _

The boiling point was again determined, and was found to be 58°7
at a pressure of 759°1 mm. The melting point was also directly
observed twice, it was —7°1°.

On considering these results, we think that the evidence goes
against the first sample, inasmuch as the second sample, after purifi-
cation, had a boiling point identical with that of the third.

The results of the vapour-pressures of the third sample are repre-
sented as a curve (No. 1), and it will be seen that the curve repre-
senting the vapour-pressures of the solid is not continuous with that
of the liquid, and that the two curves intersect at a point correspond-
ing to a temperature of —7°1°, and a pressure of 44°5 mm.

The ratios of the absolute temperatures of bromine and water at
equal vapour-pressures were calculated, and are given in the following
table. The value of c in the formula R’' = R + e (¢' — 2) is for liquid
bromine —0°000586, and for solid bromine + 0°00114. (See Phil.
Mag., 1886, p. 33.)

458 RAMSAY AND YOUNG: VAPUOUR-PRESSURES

P Water
Ratios of Absolute Temperatures of oe «
Pressure. Temp. of water. Temp. of bromine. Ratio.
Solid .... 20mm. 273 + 22°3° 273 — 16°65° 11519
» eves OF @ » 26°0 w 140 11544
n coce OW a » 29°1 ” 120 11575
~ cose &@ wp - 318 - 10°05 11592
~ «ore @ 2 342 ” 84 1°1610
~~ sooo @ - 36°3 9 70 1:1628
Liquid .. 50 ,, ” 383 - 5°05 1/1618
= soos WD a ” 517 » + 820 11547
» soce MO 9 60°1 n 16°95 1°1488
o vere BO a o” 66°6 - 23°45 1°1456
w soco OP @ - 75°9 9” 33°05 1:1400
» seco M0 o 83:0 99 40°45 1°1357 |
n wooo OD 9” 88°7 "9 46°8 11310
» voce CO 9 93°5 a 51°95 11278
s soce Oe « 99 97°7 - 56°3 11257
os toce OP 9 » 1000 - 58°75 11243

It will be seen that these ratios, mapped against temperatures of
bromine (see Curve 2) give points in a straight line; in other words c
is constant.

Vapour-pressures of Fodine.

Series I.
Solid. Liquid.
Temperature. Pressure. Temperature. Pressure.
58°1° 4°9 mm 114°1° 89°8 mm.
64°5 6°05 ,, 114°9 93°55,
66°3 6°25 ,, 117°8 1030 ,,
75 °2 115 sg 120 °4 113°4_—,,
80 *4 15°15 _,, 123 °15 124°5 ,,
86 °0 21°25 ,, 125°5 135°8 ss,
91°8 28°95 ,, 127°1 1429,
91°9 296 , 166 °6 475°0__,,
96°8 37°8 Ss, 169 *4 505°5 sy,
102°7 50°65, 171°7 535°6
105 °7 59°85, 174°5 575°3 i,
113°8 87:0, 177°6 630°3
— _— 177 °7 633°9 ,,
— _ 180 °75 680°5 ss,
— —_ (186 °4 764°2—y,
_- — | 185 *2 —
-- — + 185 °45 —
-- — 185 *45 ~_
— _ 185°55 —_
|

Journ. Chem.Soc. July 1886

Ratio of Absolute Temperatures.

1-/T0
——“60
|
|
r
——
lao
|
|
—
T
a

1-160

oo a eee ore

7-150

1140

qunuo.g jo sarmynsrduay

Journ. Chem.Soc.JSuly 1886.

° ° e c 6 ° ° ° .
S S > °
N & 8 = 8 8 SS S $s g 2 8 8
T T T T | | y 8
|
| | | |
=— +——— ~ -— + 4 + } =s — - _ —~+ = _ _4+—_—_-___+4 ——_—-+ pp 7 --—
4 | | | |
kh -+— aoe Set an ee eee Be | 8
—_ | + —~ + + + --—— -+ + + -
| | |
| | |
8 + a: An iene aeons Me Gee | + 7 + | 8
poooe “a ‘=e - | a oO + ——t + 1
i) | |
& i 8
| | |
= i
|
|
J r :
} ]
| |
3 | | | 8
= | |
| |
| t.— - A, | ~ 4
» °
&|
| —-—+ ——++—— “ } — 4 + — + —-+4+--— a 4
3} 3
|
—o + + t +
| |
st + + ! §
= — ~ 4 —= + + + + 4 - +——_— + + — —— ah 4
| 5S é
= |
2 | | } }
ws ++
a 3 | 8
—
| |
= - _ $$ —$—$—$$_$___—_— = —_——— I T p —___}_ + EE + EEE
| | |
s i Ss } x
S$ —+
J S| | 1
% | : |
1 ‘s | : +
| | |
]

S
= — :
— + | 4 + 2

}
| -
S 1 | | q : S
| |
|_| | | | | | |
| } | } j |
| | | |
| | |
8 1 + : 48
| | |
| 4 1 | + |
|
SS |
: + | +18
| |
| |
3 : .
| | | |
|
EE = | | 4 t i le
: = : : : > S S : °s ‘ : >
= AS Ss 3 s Ss & > SS 2 3 SS 8 5

gurpoy jo saxrqosaduray

Pressure uv Millimeters

OF BROMINE AND IODINE, 459

Series IT.
Solid. Liquid.
Temperature. Pressure. Temperature. Pressure.
105 *6° 61°8 mm. 127 °4° 143°1 mm.
108 °4 706 ,, 130°1 157°3 ,,
111°6 i 134°4 180°5 ,,
114°2 905 ,, 147°8 2732 ,,
antes — q 158 *4 374°6 ,,
-- — 158 °9 376°0 ,,
_ — 167 *2 475°6 ,,
— _ 174°4 5743 ,,
— —_ 177°9 625°8 ,,
= — 180°8 679°1 ,,
Liquid.

Temp. Pressure. Temp. Pressure. Temp. Pressure.
134°5° 179°9 mm. 153 °6° 323°0 mm. 163 *2° 425°9 mm.
138 °6 211°7 ;, 156 °2 340°8 ,, 165 °8 456°9 ,,
141°1 227°1 ,, 156°5 3442 ,,. 168 °0 486°6 ,,
143°0 241°0 ,, 156 °6 316°1 ,, 168 °2 488°6 ,,
147°7 273 0 ,, 158°9 369°1 ,, 171°9 536°3 ,,
150 °2 300°4 ,, 159°6 385°2 ,, 174 °2 581°6_,,
150 °7 302°4 ,, 160 *4 395°8 _,, 175°0 5838 ,,
151 °6 305°3 Co, 160 *7 399°0 ,, 175 °3: 586 °8,,
153°2 321°2 ,,. _ — _ —

These numbers are not so accordant as usual. The reason for this
is that the stem of the thermometer being invisible through the iodine
vapour, the vapour was always kept below the level of the mercury
in the thermometer, and hence a correction, not always: certain, had to
be introduced for the expansion of the mercury in the stem. It should
be stated that the temperatures of both bromine and iodine are those
of an air thermometer, for the thermometer used was calibrated by a
measurement of the vapour-pressure of water, and of the substances
described in the Transactions, 47, 640.

It may be noticed in this case also, that the curve expressing the
relation of pressure to temperature for solid iodine is not a continua-
tion of that for liquid iodine, but that they intersect at 114°3° and a
pressure of 91mm. This has been already experimentally proved
by us for a number of substances (Phil. Trans., 1884, Part II, 461).

4€0 RAMSAY AND YOUNG: VAPOUR-PRESSURES

The melting point was directly observed to be between 113°8° and
114°1°; and another observation gave 114'2°.

, Iodine
Ratios of Absolute Temperatures of Water.
Pressure. Temp. of water. Temp. of iodine. Ratio.

Solid 20mm. 273 + 22°0° 273 + 85°0° 12123
* 30 ,, - 29°1 ” 92-2 1°2089
9 50 ,, - 38°3 » 102°15 12051
- - 44°6 5, 109°05 1°2029
- 90 ,, - 49°55 » 1415 1°2003
Liquid 100 ,, ~ 51:7 » 1170 12011
- 150 ,, - 60°1 » 1289 1°'2066
” 200 ,, - 66°6 » 4137°05 1°2075
- 300 ,, - 75°9 » 1507 12144
400 ,, - 83:0 » 160°9 12188
500 ,, - 88°7 » 169°05 1°2222
600 .,, - 93°5 » 1760 1°2251
- 700 ,, * 97°7 » 182°0 1:2274
* 760 ,, ” 100°0 » 1853 12287

These points were mapped in the ordinary way. The value of c for
liquid iodine is + 0°0003986, and for the solid —0°000407. The two
lines intersect one.another at 114°3°, the melting point of iodine.

We have also constructed a diagram to show the ratios between the
absolute temperatures at constant pressures of —

1. Liquid iodine + liquid bromine, where c = —0-000080.

2. Solid — - where ¢c = —0°001125.

. “ » @~ solid - where c = +0°000413.

In these cases the ordinates are the temperatures of iodine. The
first two lines intersect at the melting point of iodine, and the second
at a temperature of iodine, where iodine vapour exerts a pressure
equal to that of bromine at its melting point. The points do not fall
well in a straight line, owing probably to the difficulty we experienced
in measuring the vapour-pressures of iodine.

From our results, the following data for the melting point of bromine
are obtained::—

1. Highest temperature attained on solidification —7°45°.

.2. Constant temperature during fusion, —7°0° and —7°05°, and with
the third sample, —7°1°.

‘3. Intersection of vapour-pressure curves of solid and liquid,
—7°05°.

4. Intersection of lines representing ratios for bromine and water,
—7°0°.

Journ. Chem.Soc.Suly ISS6

Ratios of Absolute Temperatures.

lodine

Water.

Bromine.

Sy s| | ent

- | x yo |

} oN |

q c Q om me te (ae ae See

2 8 . s = ” 2 SN 2 3 y 8

eurpol jo sarmynsaduyy,

THO

2 1-230 7380

122

1210

1200

OF BROMINE AND IODINE. 461

The first result, —7°45°, is of little value, inasmuch as the highest
possible temperature is not necessarily attained on solidification. We
believe that —7°05° expresses the true melting point very closely.

As regards the boiling point of bromine at normal pressure, the
corrected number found by us is 58°0° for the first sample; that of
the second and third samples is 58°7°; these numbers approximate
to these given by Andrews and by Landolt, 58°0° and 586° respec-
tively ; but that which we place most confidence in, 58°7°, differs from
that found by Thorpe by 0°55°, and yet Thorpe’s method of purifica-
tion appears unexceptionable. Thorpe’s remark in his paper (Chem.
Soc. J., 3'7, 172) on the thermal expansion of various bodies, that the
high solidifying point of the sample of bromine, viz., —7°5° to —8°, used
by Pierre, points to its contamination with water is, however, quite
unfounded on fact. The temperature to which a liquid can be cooled
without freezing depends on various circumstances, and is not a fixed
point ; whereas the melting point, or the point of intersection of the
vapour-pressure curves, is constant.

The following data are available for the melting point of iodine :—

1. Direct observatien, 114°1—1i4°3°.
eo - 113°8—114'1°.
3. Intersection of vapour-pressure curves of solid and liquid,

114°3°.

4. Intersection of lines representing ratios for iodine and water,
114°3°.

5. Intersection of lines representing ratios for bromine and iodine,
114°05°.

The true melting point may be expressed by the number 114°15°,
which agrees well with that given by Stas, 113—115°, and with
Regnault, who found the temperature during solidification to be
113-6°.

The boiling point of iodine at 760 mm. pressure was found to be as
a mean of five direct observations, 184°61° at 764-2 mm.; corrected
to 760 mm. the temperature is 184°35°.

It was hoped by a study of the vapour-pressures of iodine mono-
chloride to obtain results interesting on account of their comparison
with the constants of bromine. Iodine monochloride is, however, not
a stable substance. On distilling it at normal pressure, it dissociates,
viving a sublimate containing trichloride, and while most of it boils
at a temperature of 102°, the last portions boil as high as 106°, owing
to the presence of freeiodine. The vapour-pressure curve, therefore,
obtained by our usual method presents great irregularity ; the iodine
monochloride fractionates during measurement, and it is a matter of
chance if two observations coincide. We spent a great deal of time over

462 JAPP AND WYNNE: ACTION OF ALDEHYDES

this compound, and from the results obtained a curve was constructed
fitting the observed points as accurately as we were able to draw it.
On comparing the ratios between the absolute temperatures of iodine
monochloride and water, it was evident that the lines representing the
ratios of vapour-pressure of the solid should intersect that repre-
senting the ratios of the liquid ; but the results were so irregular that
the point of intersection could not be observed. All that we can say
is that the value of c for the solid body is positive, whilst it is negative
for the liquid.

Note.—Since writing this paper, a few measurements of the vapour-
pressures of solid iodine have been given by v. Richter (Ber., 1886,
1060). His results corroborate ours in a remarkable manner. At
20 mm. pressure he found a temperature of 85°; at 30 mm., 90° ; and
90 mm. at its melting point. Our nnmbers at these pressures are
85° and 922°, and at the melting point we observed 89°8 mm. pres-
sure. V. Richter, however, appears to be unaware of the fact that
we long ago made careful measurements of results which he gives for
the most part only qualitatively.

University College, Bristol.

XLVI.—On the Action of Aldehydes and Ammonia on Benzil
(continued).

By Francis R. Jarp, F.R.S., and W. Patuer Wywne, B.Sc.

Introduction.

Te reactions of benzil with aldehydes and ammonia have as yet been
studied only in the case of aldehydes of the benzene series, and the
same holds good with regard to the other diketones which have been
substituted for benzil in these reactions—phenanthraquinone and
chrysoquinone. The reactions of another dicarbonyl-compound
(glyoxal) with aldehydes and ammonia have, however, been studied
in the case of fatty aldehydes by Radziszewski, who thus effected a
new synthesis of homologues of glyoxaline.

In the present communication, we deseribe the reactions of benzil
with fatty aldehydes and ammonia. We have also in one case
employed a benzene-fatty aldehyde, cinnamaldehyde.

With fatty aldehydes, the reaction took place according to Equa-

AND AMMONIA ON BENZIL. 463

tion II (Trans., 1884, 672) of the general reactions of this class, and
substituted glyoxalines were obtained. Thus with acetaldehyde :—

C,H;-CO
| + CH,CHO + 2NH, = SC-CH, +3H.0.
C.H;-CO (

Methyldiphenylglyoxaline.

A similar reaction occurred with isovaleraldehyde. (inanthalde-
hyde, on the other hand, did not react with benzil and ammonia
under the conditions of our experiment.

With cinnamaldehyde, two distinct reactions occur. A part of the
substance reacts according te Equation III of the general reactions
(loc. cit., p. 673), thus :—

C,H;-CO
| -- 2C,H;CH:CH-CHO + 2NH,
C;H;"CO
_, sH;;CH: CH-CH-NH: CO C,H;
| + 2H,0.
~ © H,;-CH : CH:CH:NH:CO-C,H,

Dibenzoyldicinnylenediamine.

This compound is analogous to that obtained from benzil, salicyl-

aldehyde, and ammonia (loc. cit.), but unlike the salicylaldehyde
compound, could not be converted into the non-benzoylated deriva-
tive. When heated with a solution of caustic potash in methyl
alcohol, it parts with 1 mol. of benzoic acid, and is converted into an
anhydro-base :—

C,H,;-CH ; CH:CH—NH

| DO-Oes-
C,.H;CH :CH-CH——-N
Benzenyldicinnylenediamine.
This base, so far as we know, is the only compeund yet prepared
which contains the hydrogenated glyoxaline-complex—

is, however, assumed to exist in amarine.
Another portion of the cinnamaldehyde above referred to reacts

with benzil and ammonia in molecular proportions different from any
hitherto observed in this class of reactions :—

464 JAPP AND WYNNE: ACTION OF ALDEHYDES

C,.H;CO
2 | + C,H;CH:CH-CHO + 2NH; = Cy,HyN.O; + 2H,0.

C,H,-CO

Cinnimabenzil.

This, therefore, constitutes the fourth of the distinct reactions (cf.
loc. cit.) in which aldehydes and dicarbonyl compounds jointly form
condensation-products with ammonia. We shall show in a further
communication (see following paper) that a similar reaction occurs
when ammonia acts on benzil alone to form imabenzil,—nascent benz-
aldehyde, formed by the breaking up of a portion of the benzil,
taking the part of the cinnamaldehyde in the foregoing equation.
The reaction is probably, therefore, of more general application. The
foregoing analogy has led us to name the compound C,H ,N,0;
cinnimabenzil.

By treatment in the cold with a solution of potash in methyl] alcohol,
cinnimabenzil is decomposed, according to the equation—

C;,Hy»N,O oa KOH — Cy»H..N,0, + CH;'COOK.

Cinnimabenzil. Cinnidimabenzil.

When boiled with dilute sulphuric acid, it is decomposed as
follows :—

CyHy»N.O; + 2H,0 = C.H,NO, + C,H,O + C,H,O. + NH;.

Cinnimabenzil. Benzilimide. Cinnam- Benzoic
aldehyde. acid.

I. Acetaldehyde and Ammonia with Benzil.

20 grams of benzil and 8'5 grams of aldehyde were dissolved in a
quantity of alcohol sufficient to keep the benzil in solution at 40°, and
the warm liquid was then saturated with gaseous ammonia, after
which the whole was allowed to remain for 24 hours. As nothing
had separated at the end of that time, the solution was poured into
water. A white flocculent substance was precipitated, which was
thoroughly washed with water and then boiled with hot dilute hydro-
chloric acid, in which everything dissolved with the exception of a
small quantity of resin. If the acid is too concentrated, the hydro-
chloride of the new base melts to an oil under the liquid, instead of
dissolving. The base was precipitated by ammonia, and recrystal-
lised from boiling benzene, which deposited it in tufts of minute
colourless needles, with a constant melting point of 235°. It is
practically insoluble in water, but dissolves readily in alcohol and
inether. From the solution in hot alcohol, it is deposited on cooling
in well-formed crystals of the orthorhombic system (vide infra). These
had a very faint yellowish tinge.

AND AMMONIA ON BENZIL. 465

Analysis of a sample crystallised from benzene gave numbers
agreeing with the formula C\,H,N, :—

Substance, CO.. H,0.
0-1138 0°3408 0:0628
01177 0°3528 0°0640

IJI. 0:1076 gram, burnt with copper oxide in a vacuum, gave
21°91 c.c. of a mixture of nitrogen and nitric oxide, measured dry at
18°, and under 3915 mm. pressure. After absorption of the nitric
oxide there remained 21°91 c.c. of dry nitrogen at 17°8°, and under
565 mm. pressure.

Calculated for Found.
Cy6H4N3.
> atta ‘
Cy... 192 2°05 81°67 81°75
14 5°98 6:13 604
Bye. & 11:97 — — 11°95

234 10000

The constitutional formula of this compound, and the equation
P
expressing its formation, have already been given. In accordance

with this constitution, it would receive the name methyldiphenylgly-
ozaline :—

Mr. L. Fletcher has had the kindness to measure the crystals depo-
sited from alcohol. He reports as follows :—

** Methyldiphenylglyoxaline.
CoLour—Faintly yellow.

466 JAPP AND WYNNE: ACTION OF ALDEHYDES

One crystal, about a millimetre in diameter, was measured.
Systrm—Orthorhombic.
Etements—a : 6b: c:: 1°207: 1: 1°929,

or 011-010 = 27° 24 : 101-001 = 57° 582’ : 110-100 = 50° 21’.
Forms present—{010}{001}{012} {201} {111}{112}{113}.
DeveLopmMent—Vide figure.

Angles. Observed. Caiculated.
0017113 39° 55’—40° 23' 39° 52’
113°112 11° 16’, 11° 46’ ll 32
112°111 16° 11’, 16° 19’ 16
111-111 72° 27’ 72
111:010 43° 40’—44° 16’

111°012 39° 55’—40° 19’

012-113 26° 18’, 26° 25’

001-012 45° 51', 43° 56’

001°201 72° 22’—73° 32’

012-201 77° 27'—78° 2’

012°112 29° 54’, 29° 56’ 9 5:
111-201 47° 46', 47° 49° -

Methyldiphenylglyoxaline is a monacid base. In a solution of the
hydrochloride, platinie chloride (freed from excess of acid, which
might otherwise cause a separation of hydrochloride) produces a
yellow precipitate, consisting of microseopic needles of the formula
(CyeH,,N2,HCl).PtCl + 2H,0.

15446 grams of the air-dried salt lost on heating at 120°
0°0583 gram, and the resulting 1°4863 grams of anhydrous salt gave
on ignition 03279 gram platinum.

Calculated for
(Cj,H,,N,HCl).PtCl, + 2H,0. . Found.
H,O in 100 parts........ 3°94 3°77
Calculated for
(C\g¢H,4No,HCl)gPtC),. Found.
Pt in 100 parts........ 22°17 22-06

The hydrochloride itself exhibits a peculiarity in its crystallisation,
due probably to the formation of different hydrates. When the hot
saturated solution is allowed to cool in am open vessel, the crystallisa-
tion starts from ‘the surface of the solution at two or three points and
proceeds in tufts of long silky needles, extending through the entire
liquid. But if, on the other hand, the hot saturated liquid is corked
up ina vessel so as to exclude the air, taking care that none of the
needle-shaped crystals are left adhering to the cork or to the sides of
the vessel, then the substance is deposited in granular crystals, which

AND AMMONTA ON BENZIL. 467

begin growing within the liquid, and sometimes attain a considerable
size. Neither long standing nor violent shaking suffices to start the
crystallisation of needles in the closed vessel; but on removing the
cork, this crystallisation of needles commences from the surface,
showing that the liquid was still supersaturated with this form, and
wherever the needles touch the transparent granular crystals, these
become opaque, and are converted into aggregates of needles.

The behaviour of this salt recalls that of a solution of sodium sul-.
phate saturated at the temperature of maximum solubility, and pro-
tected from the air.

II. Isovaleraldehyde and Ammonia with Benzil.

An alcoholic solution of 20 grams of benzil with 8°2 grams of
isovaleraldehyde (boiling point 92—95°) was saturated with gaseous
ammonia at 40°, and allowed to stand for 24 hours. The new com-
pound separated in needles, and was filtered off. The mother-liquor
on evaporation yielded a further quantity, together with some
benzilimide (m. p. 139°5°; verified also by conversion into benzilam,
m. p. 114°). It was purified by recrystallisation from benzene, which
deposited it im silky needles, melting constantly at 223°; from hot
alcohol, it separated on cooling in larger needles.

The analytical results agreed with the formula C\.H»N, :—

Substance. CO . H,0.
0°1123 0°3396 0°0771
0°1058 0°3196 0:0720

IIT. 0°1130 gram, burnt with copper oxide in a vacuum, gave
21°91 c.c. of a mixture of nitrogen and nitric oxide, measured dry at
17°5° and under 352 mm. pressure. After absorption of the nitric
oxide there remained 21°91 c.c. of dry nitrogen at 17°5°, and under
323 mm. pressure.

Calculated for Found.

Cig oN >. c =
os I II. Ill.

a. .

228 82°61 82°47 82:38 —

Hy»... 20 7°25 7°62 7°56 —
N,.... 28 10°14 — _ 10°16

276 100-00
The compound is isobutyldiphenylglyoxaline :—

>C-CH,CH(CH,):.

468 JAPP AND WYNNE: ACTION OF ALDEHYDES

The platinichloride,* (Cjg.H»N2,HCl),PtCl,, was prepared by precipi-
tation in the ordinary way, avoiding, however, the presence of free
hydrochloric acid. It was thus obtained as an amorphous yellowish-
brown precipitate, but small crystals separated from the mother-
liquor on standing. The salt was anhydrous.

1:5956 grams dried at 100° gave on ignition 0°3222 gram platinum.

Calculated for
(C\9 HagNo, HCl). PtCl,. Found.

Pt in 100 parts 20°19

III. Cinnamaldehyde and Ammonia with Benzil.

100 grams of benzil were dissolved in sufficient alcohol to keep the
benzil in solution at 40°, 63 grams of cinnamaldehyde were added,
and the whole was saturated with gaseous ammonia at 40°. While
the gas was being passed in, a pulverulent white substance separated
in considerable quantity, and was filtered off from the warm solution
as soon as the saturation was complete. This compound was found
to be practicaliy insoluble in alcohol. The ammoniacal filtrate, on
standing, deposited a large quantity of a flocculent white compound,
which was soluble in alcohol. The filtrate from this compound yielded
nothing crystallisable on evaporation.

Ezamination of the Compound Insoluble in Alcohol.-It was found
that this compound dissolved readily im hot phenol, and was precipi-
tated on the addition of three times the volume of alcohol as a crystal-
line powder, consisting of short microscopic prisms. After repeating
this process of solution and precipitation several times, it was
obtained with a constant melting point of 264°. The compound was
boiled with successive portions of alcohel to remove the last trace of
phenol and dried finally at 120°.

Analyses agreed with the formula C;,HN,O, :—

Substance. CO. H,0.
0°1342 0°3991 0°0730
0°1250 0°3720 0°0686
III. 0°1120 gram, burnt with copper oxide in a vacuum, gave
13°07 c.c. of a mixture of nitrogen and nitric oxide, measured dry at
17° and under 351°5 mm. pressure. After absorption of the nitric
oxide, there remained 13°07 c.c. of dry nitrogen at 17° and under
318°5 mm. pressure.

* See footnote, Trans., 1884, p. 676.

AND AMMONIA ON BENZIL, 469

Calculated for Found.
C3pH og 20y. cr - >

— ~ I. II. ITI.
Cy... 384 81°35 81°11 81°16 —
Hos... 28 5°93 6°04 6:09 —
Diséee 28 5°93 = — 6:08
32 6°79 — — —

472 100°00

The yield of pure substance was 11 grams from 100 grams of
benzil, but was greatly increased in a subsequent experiment by
employing 2 molecular proportions of cinnamaldehyde to 1 of benzil.

The reaction in which this substance, dibenzoyldicinnylenediamine,
is formed, and its constitutional formula, are given on p. 463. The
action of various reagents on this compound was tried, but only with
potash was a definite result obtained.

Action of Potash on Dibenzoyldicinnylenediamine.—When the com-
pound was heated with a 10 per cent. solution of caustic potash in
methyl alcohol in a sealed tube for three hours at 150°, it went into
solution, and, on cooling, the liquid was filled with lustrous lamin, pro-
bably potassium benzoate. On opening the tube, a smell of ammonia
was perceptible. The contents of the tube were evaporated to a
semi-solid state on a water-bath, mixed with water and filtered hot.
The filtrate, on acidifying, gave benzoic acid. The brownish residue
on the filter was dissolved in hot benzene, from which it was obtained,
after two crystallisations, in small faintly yellow crystals with a con-
stant melting point of 207°.

The results of analysis agreed with the formula C,;H,.N, :—

Substance. CO. H,0.
Bossecese 0°1246 0°3898 0°0763
EE, wcccece 0°1317 0°4116 0°0896

III. 0°1085 gram, burnt with copper oxide in a vacuum, gave
13°79 c.c. of dry nitrogen at 23° and under 413°5 mm. pressure.
Nitric oxide was not present.

Calculated for Found.
Co5H92No. c _— ~
o— ee I. Il. III.
Cu---- 300 85°71 85°32 85°24 —
ae 22 6°29 6°80 6°80 ——
Pasede 28 8:00 = = 8°02
350 100°00

The compound, which we have named benzenyldicinnylenediamine
(see p. 463), is formed from the dibenzoyl-compound by the abstrac-
VOL. XLIX. 2k

470 JAPP AND WYNNE: ACTION OF ALDEHYDES

tion of 1 mol. of benzoic acid. It may be regarded as an anhydro-
base derived from the hypothetical monobenzoyldicinnylenediamine—

‘C,H;

C.H;CH : rn

as
C.H;CH : CH-CH:N|H,

by the abstraction of the elements of water.

When this base is boiled with hydrochloric acid, it is converted
into a hydrochloride, but dissolves only very slightly. The hydro-
chloride is, however, soluble in alcohol. On the addition of aqueous
platinic chloride to the alcoholic solution, the platinichloride sepa-
rates in yellow silky needles of the formula (C,;H.N.,HCl).PtCl,
+ 2H,0. The salt thus prepared was washed with alcohol, air-
dried, and analysed.

09084 gram of the air-dried salt lost, on heating at 120°, 0°0282
gram, and the resulting 0°8802 gram of anhydrous salt gave, on igni-
tion, 0°1540 gram of platinum.

Calculated for
(C.;Ha:N2,HCl)oPtCl, + 2H,O. Found.

H,0 in 100 parts........ 3°14 3°10
Calculated for

(C.;Ho2N>,HCl) 2PtCl, . Found.

Pt in 100 parts........ 17°53 17°50

Examination of the Compound soluble in Alcohel.—The flocculent
white compound (p. 468) from the benzil-cinnamaldehyde-ammonia
reaction was dissolved in boiling alcohol. The solution deposited two
sorts of crystals—slender needles and minute short prisms—which,
however, proved to be merely different crystalline forms of the same
substance, and could be converted one into the other by slightly
varying the conditions of crystallisation. Both melted at the same
temperature, 188°, and gave the same figures on analysis. Neither
contained alcohol of crystallisation. The compound is also soluble in
boiling benzene. The yield, using the proportions given on p. 468,
is about one-tenth of the weight of benzil taken.

The analyses agreed with the formula C;,HyN,0; :—

Substance. CO. H,0.
Ess vcsice 0°1252 0°3701 0°0645
Eh. ccevers 0°1326 0°3918 00675

III. 0°1074 gram, burnt with copper oxide in a vacuum, gave
13°07 c.c. of a mixture of nitrogen and nitric oxide, measured dry at
17° C. and under 2845 mm. pressure. After absorption of the nitric

AND AMMONIA ON BENZIL. A471

oxide, there remained 13°07 c.c. of dry nitrogen at 17°, and under
269°5 mm. pressure.

IV. 0°1193 gram gave 13°07 c.c. dry nitrogen and nitric oxide at
17°4° and under 325 mm. pressure, and after absorption of the nitric
oxide, 13°07 c.c. dry nitrogen at 17°2° and under 297°5 mm. pressure.

Calculated for Found.

C37H3oN 203. la - ~

int 3 II. III. IV.

Cy... 444 80°73 80°62 80°58 —_ ae

a 30 5°45 572 565 — —

Mbsawe 28 5°09 — = 5°24 5°30

Doce ss 48 8°73 —- — — —
550 100°00

Analyses I and I1I were made with the prismatic crystals, II and
IV with the needles.

The equation for the formation of the compound C;;H,N,0;—which
for reasons already stated we propose to name cinnimabenzil—is given
on p. 464.

Action of Potash on Cinnimabenzil._—The finely-powdered compound
was mixed with a 10 per cent. solution of caustic potash in methyl
alcohol, and allowed to stand in a corked flask over night. The
mixture became semi-solid, and on opening the flask a smell of
ammonia was perceptible. Water was added, and the insoluble
matter was filtered off. On acidifying the filtrate, it yielded only
benzoic acid (identified by its melting point) ; but a faint cinnamon-
like smell was perceptible during evaporation. The substance sepa-
rated by filtration was shaken with ether; this removed a small
quantity of a yellow gum, leaving the insoluble portion quite white.
The substance thus obtained was practically insoluble in the
ordinary organic solvents, but dissolved in hot phenol, and was pre-
cipitated from this solution as a crystalline powder on the addition of
alcohol. By repeating this treatment three times, it was obtained
with the constant melting point of 283°.

Analyses agreed with the formula C3H2.N,0. :—

Substance. CO,. H,0.
Riadesnes 0°1596 0°4691 0:0848
ee 0°1547 0°4547 0°0824

III. 0°1140 gram, burnt with copper oxide in a vacuum, gave
13°79 c.c. of dry nitrogen at 235° and under 339 mm. pressure.
Nitric oxide was not present.

2x2

472 JAPP AND WYNNE: ACTION OF ALDEHYDES ON BENZIL.

Calculated for Found.
C3pHogN 20>. / _ ~
ooo I. IT. II.
Cuce-- 360 80°72 80°16 80°16 —
Ho, ... 26 5°83 5°90 5°92 a
No... 28 6°28 oe —_ 6°24
Ghesee 32 7°17 — _ —
446 100°00

The substance contained a small quantity of ash which was weighed
and deducted. The mode of formation of this compound has already
been discussed (p. 464). In order to indicate its derivation from
cinnimabenzil, we have named it cinnidimabenzil.

Action of Sulphuric Acid on Cinnimabenzil.—By boiling cinnima-
benzil for a short time with dilute sulphuric acid (2 vols. of acid with
3 vols. of water), it is decomposed, yielding benzilimide, cinnamalde-
hyde, benzoic acid, and ammonia. Benzoic acid and benzilimide were
identified by the melting point, the latter also by conversion into
benzilam ; the cinnamaldehyde was distilled with steam, and obtained
as an oil with the characteristic odour. The reaction appears to take
place according to the equation—

C,H NO; + 2H,0 = C.H,,NO, + C,H,O + C,H,O. + NH.
Cinnimabenzil. Benzilimide. Cinnam- Benzoic
aldehyde. acid.

The action of concentrated sulphuric acid on imabenzil (see following
paper) corresponds with this, except that benzaldehyde is eliminated
instead of cinnamaldehyde, and the benzilimide is converted by the
concentrated acid into benzilam (C,,H,NO). Both the mode of for-
mation and the decomposition of cinnimabenzil, therefore, countenance
the view that this compound is an imabenzil in which a benzaldehyde
residue has been replaced by a cinnamaldehyde residue.

Various other reactions were tried with cinnimabenzil, but none of
them yielded definite results.

Normal School of Science,
South Kensington.

XLVII.—On Imabenzil.

By Francis R. Japp, F.R.S., and W. Patmer Wrwne, B.Sc.

THE action of an alcoholic solution of ammonia on benzil was first
studied by Laurent (Revue Scientif., 10, 122; 19, 440; also J. pr.
Chem., 35, 461), who obtained three compounds, to which he assigned
the names and formule—

Imabenzil eeereere C,,H,,NO,
Benzilimide ...... C,,H,NO,
Benzilam ......... C,,H,N.

Working under somewhat different conditions, Zinin (Awnnalen,
34, 190) prepared a compound which he named azobenzil, assigning
to it the formula C,.H3)N.O2.

Henius (Inaugural Dissertation, Marburg, 1881), at Zincke’s in-
stance, undertook an experimental revision of the work already pub-
lished on this subject. He came to the conclusion that Laurent’s
formule should be modified as follows :—

Imabenzil eeoeeesee Cy.H N20.
Benzilimide...... CywHN,0,.
Benz ilam eocccece CyHN,0, ;

and that Zinin’s azobenzil was identical with benzilam. Henius’s
results were also communicated later in a note by Zincke (Ber., 16,
890).

One of us, in a further investigation of this subject (Ber., 16, 2636),
has since shown that the formula of benzilam is C.,H,,NO, and that
benzilimide, which yields benzilam by parting with the elements of
water, is C,,H,,NO,. Part of the benzil is first broken up, yielding
benzaldehyde and benzoic acid (the latter in the form of benzamide
or ethyl benzoate), and 1 mol. of nascent benzaldehyde then reacts
with 1 mol. of benzil and 1 mol. of ammonia, forming either benzil-
imide or benzilam, according to the proportion of water eliminated
in the reaction (vide infra). The following constitutional formule
were assigned to these compounds :-—

C.HsC——O. C.HsC—O
| PCUOH) Cos, and | Je Coss
C,H;,C—NH C,H;;C—N
Benzilimide. Benzilam.

These formule have since been accepted by Henius (Annalen, 228,
339).

474 JAPP AND WYNNE ON IMABENZIL.

But as regards the third compound, imabenzil, it was admitted
(Ber., 16, 2640) that Henius’s formula, C,.H;,N,0,, was probably
correct. This opinion was, however, based upon an examination of
amorphous imabenzil prepared by the method of Henius.

We have now prepared crystallised imabenzil, and find that it
has the formula C,;H2N,0s.

Henius prepares imabenzil by adding concentrated aqueous ammo-
nia to a hot saturated alcoholic solution of benzil. A yellowish
magma is the result, which, on the addition of more ammonia,
becomes white. The whole is allowed to remain for an hour, after
which the substance is collected on a filter and washed, first with
water, then with alcohol, and finally with ether.

We find that imabenzil prepared by this method contains, mechani-
cally enclosed, unaltered benzil, which the washing with alcohol and
ether is quite incompetent to remove. This accounts for the low per-
centage of nitrogen which Henius found on analysis, and also for the
large quantity of benzil which one of us described, in the commu-
nication already referred to, as obtained by boiling the above product
with dilute sulphuric acid.

In order to obtain imabenzil, we employed the method described by
Laurent. Benzil was dissolved in alcohol, so that the solution was
saturated at 40°, and gaseous ammonia was passed into the warm
solution to saturation. After standing for 24 hours, the liquid had
deposited small colourless prismatic crystals of imabenzil. Some-
times needles of benzilimide separate along with the imabenzil, but
this can be avoided by using a more dilute solution of benzil and
allowing the ammoniacal liquid to stand longer, when well-developed
crystals of imabenzil are obtained. Benzilimide and benzilam remain
in the mother-liquor.

Henius asserts that imabenzil cannot be recrystallised without de-
composition, and he therefore analysed the amorphous white powder
obtained by his method. It is perfectly true that long boiling with
alcohol decomposes even pure imabenzil, yielding among other pro-
ducts benzilimide; but by powdering the imabenzil first, boiling only
for a short time with the alcohol, and filtering, a solution is obtained
which, on standing, slowly deposits imabenzil in perfectly homo-
geneous crystals. By employing a solvent of lower boiling point,
decomposition is avoided: the best result was obtained with methyl
alcohol. A quantity of the crude crystallised imabenzil was finely
powdered, and then boiled with a large quantity of methyl alcohol as
long as any dissolved. The solution was filtered into a flask, in which
it was allowed to stand for some days, in order to avoid evaporation
during the crystallisation. Very lustrous crystals of imabenzil, from
3 to 4 mm. in diameter, were slowly deposited. Crystallised imabenzil

Se SS

JAPP AND WYNNE ON IMABENZIL. 475

melts constantly at 194°, whereas Henius gives the melting point at
158—172°. We confirmed, however, Henius’s statement that imabenzil
heated at 140° for about an hour decomposes and melts.

The following results were obtained on analysis of the crystals
deposited from ordinary alcohol :—

Substance. CO,. H,0.
Tneeecnes 0°1273 0°3728 0°0639
| ee 0°1270 0°3725 0°0647

III. 0°1161 gram, burnt with copper oxide in a vacuum, gave
13°79 c.c. of dry nitrogen at 22°7°, and under 2945 mm. pressure.
Nitric oxide was not present.

IV. 01104 gram gave 13°79 c.c. dry nitrogen and nitric oxide at
22°5°, and under 282 mm. pressure, and after absorption of the nitric
oxide, 13°79 c.c. dry nitrogen at 22°5°, and under 277 mm. pressure.

Calculated for Found.
CysHo4N203, ‘ we :
ooo ee 8 II. III. IV.
Cy... 420 80°15 79°87 79°99 — a=
): 28 5°34. 5°58 5°66 = —
| 28 5°34 = — 5°34 5°33
er 48 9°17 — — — —
524 100°00

Analyses I, II, and III were made with one preparation; in IV a
second preparation was employed.

The values for carbon and hydrogen agree with those obtained by
Henius, whereas the nitrogen is nearly 1 per cent. higher in our
analyses. This is accounted for by the fact that the impurity con-
tained in imabenzil, prepared by Henius’s method, is chiefly benzil,
which contains C 80°00 and H 4°76 per cent. A quantity of benzil
amply sufficient to lower the nitrogen to the extent observed in
Henius’s analyses might, therefore, be present without appreciably
affecting the percentage of carbon and hydrogen.

As the point in dispute is therefore the percentage of nitrogen, we
venture to call attention to the way in which our determinations
were made. Following the method of Frankland and Armstrong,
we burned the substance with oxide of copper in a Sprengel
vacuum, and measured the nitrogen over mercury in a Frankland and
Ward’s gas-analysis apparatus, determining the quantity of nitric
oxide (if any) present, and introducing this value into the calculation.
A nitrogen determination made by this method is one of the most
exact operations in the whole of analytical chemistry. A third deter-

476 JAPP AND WYNNE ON IMABENZIL.

mination of nitrogen in imabenzil, made with a different sample, and
not included in the foregoing tabulated results, gave N 5°34 per cent.

We have to thank Mr. L. Fletcher for measurements of the crystals
deposited from methyl alcohol. He reports thus :—

“Tmabenzil.
Crystals, white.
Reach 3 or 4 millimetres in diameter.

Fie. a. Fra. 6.

. ‘une poudre crystalline qui, examinée au microscope, ne doit
offrir que des prismes droits & base rhombe, parfaitement nets, et
dont les bases sont remplacées par deux facettes triangulaires qui
reposent sur les arétes verticales obtuses du prisme.’ — Laurent’s

description.
This description applies exactly to the crystals prepared by Dr. Japp

and Mr. Wynne (see Fig. a*). Generally the prism is short as in
Fig. b, and the crystals are then octahedrid in aspect.
Owing to the character of the faces, the following elements are

only approximate :—

System—Orthorhombic.

ELEMENTS—a : b : ¢ :: 1525 : 1 : 0°788,

or 011°010 = 51° 45’ : 101:001 = 27° 20’ : 110°100 = 56° 45’.

Forms opservep—{110}{011}.
Angles. Observed. Calculated.
110°110 66° 23’'— 66° 41’ 66° 30’
110°110 113 14—113 42 113 30
011-011 78° 4' 76 30
011-011 103 8 103 30
011110 58° 54’— 59° 22’ 58 49
011°110 120 43—121 12 121 11”

* This figure also agrees with an approximate drawing given by Laurent.

JAPP AND WYNNE ON IMABENZIL.

The formation of imabenzil from benzil and ammonia may be
expressed by the equation—

3C,,H 0, + 2NH, = CysH N03 + C,H,0, + H,0O,

Benzil. Imabenzil. Benzoic
acid.

the benzoic acid being eliminated in the form of ethyl benzoate or
benzamide, substances which are always formed in the action of
alcoholic ammonia on benzil.

The formation of the three compounds obtained from ammonia and
benzil may therefore be explained by the assumption of the pre-
liminary breaking up of a portion of the benzil, yielding benzoic acid
(ethyl benzoate, benzamide) and benzaldehyde, and the condensation
of the nascent benzaldehyde with benzil and ammonia according to
the three equations :—

C,H;'CO C,H,;C——O
I. | + 0,H;CHO + NH; = | 0(OH)-C.Hs
C.H;CO C,H,.C—NH + H,0.

Benzilimide.

C,H;'CO
| + C.Hy»CHO + NH; = DC-C.H, + 2H.0.
C,H;'CO C,H,;-C—N

Benzilam.

C,;H;-CO
III. 2 | + C,H;CHO + 2NH; = C,;H,,.N.0; + 2H,0.
C,H;*CO

Imabenzil.

The last of these equations corresponds with that in which cinnima-
benzil is formed from cinnamaldehyde, benzil, and ammonia (see
preceding paper p. 464).

It is to be noted that free benzaldehyde with benzil and ammonia
yields only lophine.

Action of Sulphuric Acid on Imabenzil.—The action of cold con-
centrated sulphuric acid on imabenzil was first studied by Laurent,
who recommended this reaction as a means of preparing benzilam.
Henius finds that benzilam and benzaldehyde are formed, together
with a small quantity of benzil, but that if the solution of imabenzil
is allowed to remain for some time, or is warmed before diluting,
benzil is the chief product. We have repeated this experiment with
crystallised imabenzil, but could not succeed in obtaining benzil, and
we therefore conclude that this substance must have been contained
as an impurity in the imabenzil employed by Henius. With crys-
tallised imabenzil, we identified as the products of the reaction :

478 JAPP AND RASCHEN ON THE ACTION OF

benzilam, benzaldehyde, benzoic acid, and ammonia. The reaction
probably takes place according to the equation—

C,;H..N.0; - H,0 = C.,H,,NO + C,H,O + C,H,0, + NH;,

Imabenzil. Benzilam. Benz- Benzoic
aldehyde. acid.

and corresponds with that in which cinnimabenzil is decomposed by
boiling with dilute sulphuric acid (see preceding paper, p. 464),
except that in the present case benzilam is formed instead of benzil-
imide, owing to the action of the concentrated acid.

But by boiling finely-powdered crystallised imabenzil with dilute
sulphuric acid (1 vol. of acid to 2 vols. of water) for one hour, the
imabenzil was converted into benzilimide and benzil; apparently in
equal molecular proportions :—-

CyH..N,0,; + H,O = CyHyNO, + CyHwO, + NHs.

Imabenzil. Benzilimide. Benzil.

The proportion of benzilimide to benzil found was as 1°4: 1, the
theory demanding 1°5: 1.

Boiling with glacial acetic acid also converted imabenzil into a
mixture of benzil and benzilimide as described by Henius.

We are unable to suggest a constitutional formula for imabenzil.

Normal School of Science,
South Kensington.

XLVIII.—On the Action of Phosphoric Sulphide on Benzophenone.

By Francis R. Japp, F.R.S., and Jutius RascHen.

Tue action of the sulphides of phosphorus on ketones appears to
have been but little investigated. Louguinine and Lippmann found
that camphor, when heated with phosphoric sulphide, parts with
the elements of water, yielding cymene. Wislicenus, by acting on
acetone with phosphorous sulphide, obtained duplothiacetone (C;H.S)2,
and states that ethereal salts of phosphorous acid are formed at the
same time.

We resolved to study the action of phosphoric sulphide on benzo-
phenone. It seemed probable that in this case the action of the
sulphide would be restricted to the carbonyl-group, and that we

PHOSPHORIC SULPHIDE ON BENZOPHENONE. 479

should thus be attacking in its simplest form the problem of the
action of this sulphide of phosphorus on a ketone. We find that
two different compounds are formed, according to the temperature
employed.

The action of phosphoric sulphide on a compound closely allied to
benzophenone—benzhydrol—has however been studied by C. Engler,
who in this way obtained a substance, C,,H»S,, melting at 151°
(which he regarded either as a disulphide of the radicle (C,H;),CH-,
or as a thiopinacone) together with an oily compound which he was
unable to purify, but which formed a mercury compound agreeing in
composition with that derivable from the thioalcohol (C,H;),CH-SH.
The compound, C,.H».S2, was also obtained by the action of potassium
hydrosulphide or ammonium hydrosulphide on benzophenonidene

dichloride, (Cs;H;),CCl».

Action of Phosphoric Sulphide at 100°.

Benzophenone was mixed with twice its weight of powdered phos-
phoric sulphide and heated for about 16 hours at 100°, shaking from
time to time; the mass became green and afterwards blue. The
product of the reaction was decomposed with water, and the insoluble

residue was boiled with a large quantity of alcohol. The alcoholic
solution deposited a greyish powder which was dissolved in a small
quantity of benzene. By spontaneous evaporation, the benzene
solution deposited large crystals of two kinds—colourless and yellow
—which were separated mechanically. The yellow crystals proved to
be sulphur. The colourless crystals, which became white and opaque
on exposure to the air, were purified by recrystallisation, finally from
alcohol, which deposited the substance in lustrous flat needles, melt-
ing at 152°. It agreed in its properties with the compound C.H2.S,
described by Engler. A sulphur determination (Carius) gave the
following result :—

0°4505 gram substance gave 0°5469 gram barium sulphate.

Calculated for C.,H2.S>. Found.
S in 100 parts............ 16°08 16°57

It is deposited from benzene in two forms—either by spontaneous
evaporation in large tabular crystals of monoclinic habit, which
become opaque by exposure to the air or on heating to 100°, and
probably contain benzene of crystallisation, or, by cooling of the hot
saturated solution, in needles, which more closely resemble those
from alcohol and do not become opaque.

Heated above its melting point to 200°, it turns deep blue.

480 JAPP AND RASCHEN: ACTION OF PHOSPHORIC SULPHIDE.

Action of Phosphoric Sulphide at 140—150°.

A mixture of 50 grams of benzophenone with 100 grams powdered
phosphoric sulphide was heated at 140° in an oil-bath until a ther-
mometer placed in the mixture registered the temperature of the oil-
bath. At first the temperature inside the flask rose to 150° owing to
the heat given off in the reaction; as soon as the equalisation of tem-
perature had taken place, the action was regarded as complete. The
contents of the flask, which were of a deep blue colour, were boiled
with water to decompose the excess of phosphoric sulphide, after
which the insoluble matter was collected and boiled out twice with
an excess of alcohol, washing with boiling alcohol on the filter. The
greyish pulverulent residue was dissolved in boiling benzene, in which
it was only sparingly soluble, and the filtered solution was mixed
with twice its volume of light petroleum; this caused the separation
of a white crystalline powder consisting of microscopic elongated
plates. The substance was then recrystallised from boiling benzene,
without the aid of petroleum, until it melted constantly at 226—227°.
In melting, it decomposes, assuming a deep blue colour. It is practi-
cally insoluble in alcohol; hot benzene dissolves it sparingly, and on
cooling deposits nearly the whole of the substance in the above-men-
tioned crystalline form.

In addition to carbon and hydrogen, it contained both phosphorus
and sulphur. Analysis led to the formula CH2»P,S; :—

Substance. CO,. H,0.
0°3218 0°6669 0°1150
0°5923 0°0945

III. 0°8072 gram, oxidised by heating with sodium carbonate and
mercuric oxide, gave 0°3119 gram of magnesium pyrophosphate.

IV. 05103 gram, oxidised with nitric acid, gave 0°2006 gram of
magnesium pyrophosphate.

V. 0°5331 gram, oxidised by heating with sodium carbonate and
mercuric oxide, gave 1°1526 grams of barium sulphate.

VI. 03956 gram, oxidised with nitric acid, gave 0°8488 gram of
barium sulphate.

Calculated for Found.
Cy5HoP2S;. a A
(a, 8 Il. ° ° V.
56°32 56°52 56°26 —
3°61 3°97 3°66 —_ —
11°19 a — 10°79 10°97 —
— a — — 29°69 29°47

BAILEY: SEPARATION AND ESTIMATION OF ZIRCONIUM. 481

The substance is very hard to burn, and the full amount of carbon
can only be obtained in analysis by mixing the substance with a large
quantity of powdered lead chromate and heating very strongly.

When the compound is oxidised with a boiling solution of chromic
anhydride in glacial acetic acid, benzophenone is regenerated. No
substitution of sulphur or phosphorus in the phenyl-groups therefore
takes place in the formation of this compound.

The formation of such a compound is most readily accounted for
by supposing that thiobenzophenone is first formed by the replace-
ment of oxygen in benzophenone by sulphur, and that 2 mols. of this
compound then unite with 1 mol. of phosphorous sulphide (either
contained in the phosphoric sulphide or, more probably, formed from
it by abstraction of sulphur) to yield the new compound,

2CS(C.Hs)2 + P.S; a C.sHoP.Ss.

The most probable formula for such a compound, formed by the
union of 2 mols. of the sulphide of a dyad radicle, C(C,H;). (which
might be named benzophenonidene), with 1 mol. of the thianhydride,
P.S;—and therefore presumably an ethereal thio-salt—would be—

P<g>C(CoHi
S

|
P<S>O(C.Hs)

We propose to name this compound benzophenonidene pyrothiophos-
phite.

Normal School of Science,
South Kensington.

XLIX.—A Method for the Separation and Estimation of Zirconium.

By G. H. Bartey, D.Se., Ph.D., Assistant Lecturer in the
Owens College.

Ir has already been pointed out (this vol., p. 149) that zirconium is
completely precipitated from its solution in dilute sulphuric acid on
addition of hydrogen peroxide alone. Hermann (J. pr. Chem., 97,
331), has described elaborate processes for the estimation of zirconium,
but was unable to effect a satisfactory separation from titanium, iron,
and some of the rarer earths. It was therefore desirable, in continu-

482 BAILEY: A METHOD FOR THE

ance of the work referred to above, to ascertain whether the separa-
tion by means of hydrogen peroxide is capable of application in
presence of varying amounts of such substances. Judging from
Cléve’s results (Bull. Soc. Chim., 43, 53), there can be little doubt as
to the similarity of thorium and zirconium in relation to this reagent,
and where these two elements occur together they will in all proba-
bility both be precipitated on adding hydrogen peroxide to the solu-
tion; it is not proposed to consider the case of thorium in this place.
Since, however, niobium and tantalum frequently accompany zirco-
nium, the first has been included in the mixture. Solutions were
made of the following composition :—

(1.) Containing 0°02086 gram ZrO, per cubic centimetre.
(2.) 0°00417 .

(3.) 000040 TiO,
(4.) 0:00446 ~ =
(5.) , 000071 Nb,.O; __,,
(6.) 0°00625 Fe,0, _,,

””?

Hydrogen peroxide was added to a portion of each of these, and,
except in the case of zirconium, no precipitation occurred even after
several weeks. A mixture containing zirconia in larger quantity than
the other constituents was made for determination from solutions 1,
3, 5, and 6. To the moderately acid solution, excess of hydrogen per-
oxide was added, and after standing in a stoppered flask 24 hours, the
precipitated oxide was collected on a filter. It was perfectly white
even after ignition, and weighed 0°4115 gram; no trace of iron or
titanium could be detected in it. The filtrate was now diluted and
boiled for 30 hours, the excess of the precipitating agent was thus
decomposed, aud a large part of the niobie and titanic oxides sepa-
rated out. These were weighed together, the titanium being after-
wards determined by Weller’s colorimetric method and the niobium
by difference. The filtrate from this still contained iron and some
titanium, which were precipitated by means of ammonia. The results

Taken, Found.
0°4105 gram 0°4115 gram.

0-0103 \ 00240,

00965,

A second portion was taken containing relatively less zirconia and
being altogether more dilute than the above, the solutions 2, 4, 5, and
6 being used. The precipitation and analysis were carried out
exactly as in the previous case.

SEPARATION AND ESTIMATION OF ZIRCONIUM.

The results were—
Taken. Found.
007188 gram 0°0745 gram
001450 _,, :
005288 _,, } OOS»
0:03887 _,, 0°0377_—,,

The general results being thus satisfactory, a third solution was
made up with a view to testing the delicacy of the separation, as it is
most important to be able to determine small quantities of zirconia in
presence of a large excess of other oxides. The solution occupied
32 c.c. in bulk and contained—

The precipitation of zirconia in this case was complete only after
standing two or three days. The precipitate weighed 0:0055 gram
and was free from iron and titanium. It was not considered necessary
to determine the titanium and iron. As it is most desirable to have
a simple and trustworthy method for the complete separation of
zirconium from titanium and iron, further experiments were made in
this direction.

Zirconia was precipitated by hydrogen peroxide in the presence of a
very considerable excess of iron and of titanium respectively, the pre-
cipitate redissolved whilst still moist in dilute sulphuric acid, and the
solution very carefully tested for these two elements. In no case was
the zirconia found to contain either iron or titanium.

Properties of Zirconium Pentowide.

The oxide described in my former paper is of course a hydrated
oxide. In this form even when freshly precipitated it is insoluble in
very dilute (1 per cent.) sulphuric acid or in dilute acetic acid, though
on boiling for some time it partially dissolves, being probably first
decomposed.

The ordinary hydrated dioxide on the other hand, especially in the
freshly precipitated condition, is readily dissolved by these acids, and
indeed a separation of the two oxides can be made by treating a
mixture of the two with such acid. Oxalic acid dissolves both oxides
quite readily, and the fact may here be emphasised that zirconia in
the moist state dissolves in this acid; for although Berlin (J. pr.
Chem., 58, 147) has already pointed this out, the contrary is stated to
be the case in a large number of authoritative text-books. In dilute

484 BAILEY: A METHOD FOR THE

acetic acid, the pentoxide is quite insoluble, and it was therefore not
unlikely that the precipitation might be more advantageously carried
out in this medium than in the presence of mineral acids. A repeti-
tion of Haas’ work (Ber., 17, 2249) on the peroxides of cadmium,
substituting the acetate for the sulphate, showed that whereas the
latter gives an oxide of the composition Cd,O,, the former gives not the
faintest trace of peroxide. In the case of zirconium, however, the
oxide obtained by precipitation in an acetic acid solution proved to
be the pentoxide and gave the ratio—

ZrO, : O :: 122 : 73,
the calculated relation for the pentoxide being—

ZrO, : O :: 122: 8.
A portion of the pentoxide was now taken and dried at 100°, and the
oxygen determined by the iodometric method, the zirconia by heating
a separate portion over the blowpipe until the weight was constant.

The water of hydration was obtained by difference.
The results were—

Portions dried at 15° over sulphuric acid gave—
74°61 75°66
2°44, 3°13
22°95 21:21

The oxide seems therefore to have the composition Zr,0;,4H,0,
though in the dry state it is not of constant composition, and loses at
ordinary temperature about one-fifth of its peroxide oxygen, and at
100° about one-half.

Cléve has (loc. cit.) described an oxide containing more oxygen than
the foregoing, which he obtained by adding ammonia and hydrogen
peroxide to a salt of zirconium.

Although hitherto the several specimens of peroxide, which I have
prepared, had invariably proved to be the pentoxide, I have recently
obtained the higher oxide, and this also on addition of hydrogen per-
oxide alone.

What the conditions are under which this higher oxide is produced
I am not prepared to say. I obtained it by treating the solution of
the sulphate (prepared by heating the double fluoride of potassium
and zirconium with sulphuric acid) immediately with hydrogen

SEPARATION AND ESTIMATION ON ZIRCONIUM. 485

peroxide before the basic potassium salt had had time to separate.
In previous preparations, the solution had been allowed to stand and
was filtered before adding the peroxide; it was therefore more dilute.
This higher oxide is no doubt the compound described by Cléve, and
has the composition ZrQ3.

In its properties, it resembles the pentoxide, behaving in the same
way towards acids.

It seems, however, to be more stable, and can be dried at ordinary
temperatures without undergoing decomposition. This is shown in
the following experiments. The oxide was freed from moisture by
placing it in a stream of air dried over caustic potash and sulphuric
acid. Portions were then analysed in the same manner as the
pentoxide, and the results obtained gave the following percentage
composition :—

MAO s oc cccecs 52°17 51°84 —
7 scenesauns — 7°01 6°85
H,0 ecvcccce — 41°15 anw

This answers fairly well to the composition ZrO;,5H,O. Dried at
100°, it gave—

MGM cccscses 59°61 59°17
O eeree @rererve —- 3°41
BD. cervccose — 37°42

At this temperature, therefore, it loses about 1 mol. H,O and half
its additional oxygen. The water of hydration cannot be looked on
as perfectly definite in amount. Calculating for the anhydrous
oxides we have the following numbers :—

ZrO». oO.
Dried at 15° .....0.. 88°13 11°91
<— oonenes — 11°64
< 2 sien 94°55 5°45
Calculated for ZrO;.. 88°41 11°59
and for the pentoxide :—
ZrOx. O.
Dried at 15° ........ 96°32 3°68
—— gree. 95°07 4°93
Calculated for Zr.0,.. 93°85 6°15

Note.—In a private communication with which I have been
favoured by Prof. Cléve, he suggests that these peroxides may possibly
contain traces of sulphuric acid, but on examination I can obtain no
evidence of such contamination.

VOL. XIX.

L.—Derivatives of Taurine. Part II.

By J. Wittiam James, Ph.D., F.C.S., University College of South
Wales, Cardiff.

Tue preparation of trimethyltaurine by the action of a concentrated
alcoholic solution of trimethylamine on the corresponding salt of
8-chlorethylsulphonic acid has been already described (Trans., 1885,
372). I have since then made larger quantities of this substance, and
carried out a few experiments with a view to determine whether
trimethyltaurine is to be regarded as the methyl ether of dimethyl-

taurine, CH<S0, ‘OMe , as is warranted by its synthetical formation,

or as a compound containing pentad nitrogen, OH< 50>, analo-

gous to trimethylglycocine, CH<G0>0 (betaine).

Decomposition of Trimethyltaurine with Alkalis.

15 grams of pure trimethyltaurine were dissolved in water, 20 grams
of crystallised barium hydroxide added, and the mixture boiled in a
flask until the steam was neutral to litmus-paper. The evolved gases
were passed through a well-cooled U-tube containing dilute hydro-
chloric acid. This solution was afterwards distilled very carefully to
one-third, the distillate twice distilled to one-third, potassium carbo-
nate added, and again distilled; on testing this last distillate for
methyl alcohol, no trace of it could be detected. A portion of the
hydrochloric acid solution was then precipitated with platinum tetra-
chloride in three equal portions, and the dried platinochloride analysed
with the following results :—

Fraction I. 1. 0°2474 gram substance, after ignition, yields

0°0909 gram platinum.
2. 0°3272 gram substance, after ignition, yields
0°1204 gram platinum.

Fraction IT. 02141 gram substance, after ignition, yields

0:0790 gram platinum.
Fraction ITI. 0°243 gram substance, after ignition, yields
0089 gram platinum.

Cale. for trimethylammo- Found. Fraction.
nium platinochloride, cr A \
2N(CH;);HCI,PtCl,. I. II. III.

Pticscsses 36°81 p. c. 1. 36°74 36°89 36°62

2. 36°79 — —_

JAMES: DERIVATIVES OF TAURINE. 487

These numbers leave no doubt that the gas evolved in the decom-
position of trimethyltaurine is nothing but pure trimethylamine.

The residue in the flask containing the barium was heated to
boiling with addition of more water, and then poured into a hot
aqueous solution of 6°28 grams of concentrated sulphuric acid to
remove the barium. After filtering off the barium sulphate, the
strongly acid liquid was evaporated to a syrup on the water-bath,
diluted with water, neutralised with potassium carbonate, and evapo-
rated to dryness. The residue of crude potassium salt was well
ground up with strong alcohol and digested with it for some time; on
filtering the solution, silky needles were soon deposited which
resembled potassium isethionate both in appearance and in being
sparingly soluble in alcohol. The substance after recrystallisation was
dried at 100°, and a complete analysis has shown that it consisted of
the above salt in a state of purity.

I. 0°2054 gram substance burnt with CuO,PbCrO, and Cu gave
00590 gram H,0 and 0°1065 gram CO,.
II. 0°205 gram substance fused with KHO and KNO, produced
0°3006 gram BaSQ,.
III. 0°1786 gram substance produced 0°095 gram K.SQ,.

Calculated for Found.

potassium isethionate, r “~ ~

OH-C,H,:SO3K. ) 3 II. III.

Gievestsecs 24 14°63 14°14 —- a

Mieseeesns 5 3°05 3°19 — —

csikeoues 32 19°51 — 20°13 _—

P sscovess 64 39°03 —_— — —
Te esasscee 39 23°78 _ _ 23°84

164 100°00

From the above experiment it follows that trimethyltaurine when
boiled with excess of alkali decomposes, according to the equation—

NMe,
-SO,

into trimethylamine and isethionic acid.

C.H.< >0O + H,0 = NMe,; + OH-C,H,SO,H,

Cyanamide and Trimethyltaurine.

15 gram of trimethyltaurine and 0°37 gram cyanamide were
allowed to stand in concentrated aqueous solution for some days, but
no action took place: the mixture was afterwards heated in a sealed
tube for four hours at 120°; on evaporating the liquid, a crystalline
substance separated, insoluble in alcohol, and having a sweet taste:

212

488 JAMES: DERIVATIVES OF TAURINE.

an estimation of the nitrogen proved it to be unaltered trimethyl-
taurine.

01616 gram substance produced 12°3 c.c. moist nitrogen at 15°
and 757°7 mm.

Calculated for
trimethyltaurine. Found.

8°73

Hydriodic Acid and Trimethyltaurine.

So far as is known, compounds containing a methoxyl-group are
decomposed by boiling with hydriodic acid, sp. gr. 1°68, with forma-
tion of methyl iodide, and the number of methoxyl-groups can be
easily estimated by converting the methyl iodide into silver iodide as
described in a recent paper by Zeisel (Monats. Chem., 6, 989; Abstr.,
1886, 493). An experiment carried out precisely as described in this
memvir gave no trace of methyl iodide after boiling for an hour;
this appears to me to exclude the formula NMe,*C,H,SO,OMe
(Trans., 1885, 367), as such a compound would undoubtedly yield
methyl iodide under this treatment.

On heating trimethyltaurine with hydriodic acid in a sealed tube at
200°, a reaction takes place with separation of iodine ; a strong odour
of mercaptan compounds is produced, but no methyl iodide even
under these circumstances could be detected.

Constitution of Taurine and its Alkyl-derivatives.

The ease with which trimethyltaurine is decomposed by alkalis into
trimethylamine and isethionic acid seems to show conclusively that it
must have a constitution very different from tanrine and its deriva-
tives with monamines and diamines, for had these compounds a
similar constitution there is no reason why they should not also be
decomposed by boiling with alkalis into monamines and. diamines and
isethionic acid, but it has already been shown that such is not the
case (Trans., 1885, 368).

Again, the acid reaction of taurine, mono- and di-substituted
taurines, and their behaviour with cyanamide strongly supports the
above supposition, as all these substances unite with it forming well
characterised crystalline compounds (Engel, Ber., 8, 1597 ; Dittrich,
J. pr. Chem. [2], 18; James, Trans., 1885, 373). Trimethyltaurine,
on the other hand, is neutral to litmus, and forms no combination with
cyanamide under similar conditions.

It follows then that trimethyltaurine is to be represented as a
saturated compound containing pentad nitrogen and analogous to

JAMES: DERIVATIVES OF TAURINE, 489

trimethylglycocine (betaine), whereas taurine and mono- and di-
substituted taurines containing triad nitrogen are capable of com-
bining with a molecule of cyanamide. The following general formule
for taurine and its alkyl-derivatives,

Taurine. Monalkyltaurine. Dialkyltaurine. Trialkyltaurine.

NHR’ NR’ NR’
CH<SO20H CH<§0,0H CHK<§0,0H C<§9,!>9,

would then be in accordance with the experimental evidence, and a
trialkyl taurine, as a compound containing pentad nitrogen, would
presumably decompose more or less readily into two molecules, which
has been shown to be true of the trimethy]-derivative.

Action of Ammonia on B-Chlorethylsulphonic Chloride, Cl-C,H,ySO,Cl.

Some time ago I investigated this reaction with the intention of
preparing the amide of f-chlorethylsulphonic acid, but only suc-
ceeded in obtaining an oily substance free from chlorine (Trans.,
1883, 47), and in such small quantities that I was unable at the time
to study it further. Working now with larger quantities of pure
A-chlorethylsulphonic chloride, I have obtained satisfactory results.
30 grams of the chloride were dissolved in 300 c.c. of absolute ether
and dry ammonia gas passed in to thorough saturation, the flask being
well cooled during the experiment. The precipitate which had formed
was collected, dried over sulphuric acid, and analysed.

I. 0°398 gram substance burnt with CuO, PbCrO,, and Cu gave
0°2265 gram H,O and 0°1852 gram CO,.
II. 0°2454 gram substance burnt with CuO and copper gauze with
sodium hydrogen carbonate, produced 44 c.c. of moist nitro-
gen at 19° C. and 767 mm.
III. 0°581 gram substance dissolved in water and acidified with
nitric acid, produced 0°775 gram AgCl.
IV. 0°528 gram substance treated similarly gave 0°7025 gram

AgCl.
Calculated for

Calculated for CHC | +
NH,-C,H,-SO,NH, so,
+ 2NH,Cl. 2NH,Cl.

10°39 11-21

6°92 6°07
24°24 19°62 % a= —
30°73 33°17 — 32°99 32°89
13°85 14°95 _-_ —
13°87 14-98 — a=

100-00 100°00

JAMES: DERIVATIVES OF TAURINE,

490

From these numbers, it is clear that the precipitate consisted of a
mixture of a substance of the formula C,H;NSO, with ammonium
chloride: this is specially apparent from the nitrogen and chlorine
determinations. Its formation is expressed by the following simple
NH
equation: Cl-C,H,SO,Cl + 3NH, = 2NH,Cl + CHA |,
SO

unless indeed the compound, NH,°C,H,SO,NH, (the amide of taurine,
NH,°C,H,SO,0H), is so unstable that it decomposes on standing over
sulphuric acid in the cold, which seems hardly probable.

The ammonium chloride was easily removed from this precipitate
by adding silver oxide in slight excess, evaporatiug the mixture to
dryness on the water-bath, taking up with water, and filtering off the
silver chloride. The filtrate was found to still contain a small quantity
of silver which was got rid of by hydrogen sulphide ; on evaporating
the liquid, now free from silver, to dryness, and drying at 100°, the
new substance was obtained as a yellowish oil which, on cooling, set
to a hard mass resembling gum arabic.

An analysis (I and II) gave the following numbers, and they agree
NH

as well as can be expected with ethanesulphonimide, CHC |
(anhydrotaurine). The other nitrogen determinations (III, IV, and
V) were made from separately prepared samples.
I. 0°4604 gram substance burnt with CuO, PbCrO,, and Cu
produced 0°2312 gram H,O and 0°359 gram COQ.
II. 0°3562 gram substance burnt with CuO and copper gauze
with sodium hydrogen carbonate, gave 43 c.c. moist
nitrogen at 10° and 763 mm.
III. 0:178 gram substance similarly treated gave 21 c.c. moist
nitrogen at 16° and 765 mm.
IV. 0147 gram substance by the soda-lime method yielded
0°137 gram platinum.
V. 0113 gram substance by the soda-lime method yielded
0°110 gram platinum.

Calculated for ethanesulphonimide.

NH Found.
CHK] . ——
SO, : § EA. III. BV. ¥:
C,.... 24 2243 a =—- =
H..... 5 467 ee
rrr 13:06 — 1434 13°60 129 137

Bo scsce 32 29-92 -- — on —_: wh
oe 4 29°92 — — ma —

100°00

JAMES: DERIVATIVES OF TAURINE. 491

Ethanesulphonimide is a solid resembling gum arabic in appear-
ance: it melts between 45° and 50°, and at 100° becomes more or less
mobile. It has a bitter taste. In cold water, it is scarcely soluble,
but mixes with hot water in all proportions: it is insoluble in alcohol
or ether.

No taurine was formed on heating it with water alone in a sealed
tube at 150°, neither did the addition of a drop of hydrochloric acid
bring about a reaction. No compound with hydrochloric acid and
platinum tetrachloride could be obtained. It is not decomposed by
boiling with barium hydroxide, and this may be taken advantage of
for obtaining this substance from its admixture with ammonium
chloride, the barium being subsequently removed by sulphuric acid.

Metallic Derivatives of Ethanesulphonimide.

If freshly precipitated mercuric oxide is gradually added to a hot
concentrated aqueous solution of ethanesulphonimide, it is dissolved,
and on filtering the liquid, after a slight excess of the oxide has been
added, the mercury compound separates on cooling as an amorphous
white precipitate which aggregates into several semi-solid globules
on standing. The precipitate was washed once or twice with distilled
water and dried over sulphuric acid.

I have not been able to prepare this mercury compound in a state of
purity, it is always mixed with more or less of the original substance,
probably carried down with it mechanically, but the following num-
bers obtained on the analysis of a sample can leave no doubt but that
the hydrogen of the imide-group has been replaced by mercury—two
molecules of imide being required for one molecule of mercuric
oxide—

NH ¥ N ~
26H | + Hg0=|CHi<| | Hg + H,O.
SO, be SO, _|2
I. 0°2300 gram substance burnt with CuO and copper gauze
produced 15:6 c.c. moist nitrogen at 13° and 757 mm.
= 7°86 per cent. N.
II. 0°1185 gram substance dissolved in hydrochloric acid with
addition of a drop of nitric acid, gave 0°0548 gram HgS
= 0°04724 gram Hg = 39°86 per cent.

If the excess of nitrogen over that required by the formula of the
mercury compound be calculated to ethanesulphonimide and sub-
tracted, the percentage of mercury is as near as possible the theo-
retical.

492 JAMES: DERIVATIVES OF TAURINE.

01185 —0°0213 (admixture of imide calculated from excess of N)
= 0°0972 gram pure mercury compound.
Calculated for
(C,H,NS8O,).Hg. Found.
48°60

Silver oxide is also dissolved by the aqueous solution of ethane-
sulphonimide, and doubtless forms a compound analogous to the
above.

In the Berichte, 18, 869, Leymann describes experiments with
A-chlorethylsulphonic chloride and aniline: he obtained three sub-
stances in this reaction—

NPh

Cl-C,H,ySO,.NHPh; NHPh-C,H,SO,NHPh; and OH d
V2
I have found no indication of any analogues of the two first com-
pounds, but this ethanesulphonimide is the compound corresponding
to the latter substance, which Leymann has named anhydrophenyl-
taurine. It is true such a compound and water should yield phenyl-
taurine or phenyltaurine on hydrolysis, but there is no evidence to
support these assumptions, and I have in vain endeavoured to prepare
taurine from ethanesulphonimide by heating it with water and dilute
acid under pressure. The name anhydro is therefore misleading and
should be discontinued in connection with this new class of sub-
stances, for they are neither acids or anhydrides, but simply analogous
to the imides of the fatty series with carboxyl. Leymann’s body
should be called phenylethanesulphonimide, and the relation between
taurine and ethanesulphonimide is plainly seen from the following
formule :—

NH,’C,H,SO,-OH, amidoethanesulphonic acid (taurine).
NH,’C,H,SO.NH;,, amidoethanesulphonamide (unknown).
NH
CHK | , ethanesulphonimide.
SO.

LI.—The Influence of Remelting on the Properties of Cast Iron. Notes
on Sir W. Fairbairn’s 1853 Experiments.

By Tuomas Turner, Assoc. R.S.M., Demonstrator of Chemistry, Mason
College, Birmingham,

In a recent paper, published in the Journal of this Society (Trans.,
1886, p. 130), I have entered at some length into a consideration of
the experiments conducted by Sir William Fairbairn to ascertain the
effect of repeated remelting on the properties of cast iron (B. A. Re-
port, 1853, p. 87). I endeavoured to show “ that the effect attributed
to remelting may be sufficiently explained by a consideration of the
chemical changes which took place.” These changes I then believed
to be a gradual increase of silicon (owing to the metal being melted in
a cupola in contact with fuel), accompanied by the fact “ that sulphur
considerably increased about the fourteenth and fifteenth melting,
rendering the metal hard, white, and brittle.”

In the discussion which followed the reading of that paper, it was
stated by Professor Unwin, who assisted in these experiments, that
remelting was not conducted in a cupola, as I had been led to suppose
in the first place, but that an air-furnace was used for this purpose.
Since then a fortunate opportunity has been presented to me for the
examination of what are probably the only specimens of these
test bars now in existence. The result of this examination has been
to considerably modify my original conclusions as to the details of the
changes, though it has amply confirmed my contention that the effects
observed could be sufficiently explained by a consideration of the
changes in chemical composition.

My former conclusions were based upon the following analyses,
which were given in the original paper on the authority of Professor
Calvert :—

Percentage of

No. of meltings. ji. 8. C.
0°42 2-76
0°60 2°30
0°26 3°50
0°75 3°75

It will be shown later that these analyses are altogether erroneous,
and at present I am not able to make out any connection whatever

494 TUKNER: THE INFLUENCE OF REMELTING

between the numbers given above and the actual composition of the
specimens.

For the specimens I have been able to examine, I am indebted to the
kindness of Professor Unwin, and I would here record my deep in-
debtedness to him for resigning into my hands, for the purpose of
chemical analysis, specimens which he highly prized, and which are
of special value as the only remains of the experiments of 33 years
ago. It may be well to mention here that the identity of these pieces
is thoroughly assured. In forwarding them Professor Unwin said,
“There is absolutely no doubt that the pieces sent you were broken off
Fairbairn’s remelted bars. I have preserved absolutely no other
specimens of cast iron, so there has been no mixing of two sets. The
only possible doubt is as to the numbering.”

As a doubt might be possible in the latter respect, a short account
may be given of the pieces received. Of these there were seven in
all, each being 1 inch square in section, as described in the original
paper. The largest piece was ubout 1 inch in length also. Five of
the specimens were numbered (Nos. 1, 8, 14, 15, 18), the labels being
gummed on, and having been attached over 25 years ago. In each
case the character of the metal exactly agreed in hardness and
appearance with the description given in the original paper, which
contains a minute account of the fractures. In two of these pieces,
Nos. 15 and 18, we have a further confirmation in the sketches
given in the paper, which exactly correspond with the metal
examined. In one case a specimen was not numbered. Its appear-
ance was very characteristic, white on the outside, but with a small
circle of grey inside about three-eighths of an inch in diameter. This
appearance was sketched in the original memoir, and made it certain
that the piece was from the sixteenth melting. In one case, the label
had become quite illegible. As there was a considerable amount of
iron rust on this label, it was not possible to restore the ink marks by
ordinary processes. The label was therefore carefully detached and
held up to the light. The figure 1 was plainly seen, and part of a
second figure could also be discerned. From its shape and other
circumstances this could be only either a 2 or a 3, and the evidence
appeared to be strongly in favour of 12 as the right number. Professor
Unwin said of it, “I have a strong suspicion that this is the bar of
maximum transverse strength (12th), but of course I have no proof.

I think it is almost certain I should have retained a specimen of that

bar.” In hardness it was intermediate between 8 and 14, and its

composition also favoured this view, so that the identity of this speci-
men may be considered assured.

The chemical examination was rendered more than usually difficult,
owing to the comparatively small quantity of material to be operated

ON THE PROPERTIES OF CAST IRON. 495

upon, and the necessity for the preservation of a portion of each piece
for reference at any future time.* In one case also (the 18th melting),
the specimen was so hard that it could not be touched by a file or
drill, and a sample for analysis could only have been obtained at great
risk to the specimen itself. As Professor Unwin prized this particular
piece more highly than several of the others, and as the analyses
already made rendered the action quite evident, it was not considered
well to proceed further than an estimation of combined carbon in this
case. When it was found that my analyses gave results so very
different from those of Professor Calvert, I thought it well to forward
my results, so far as they had then gone, to Professor Unwin, suggest-
ing that samples should be sent to some independent chemist of known
experience in iron analyses. As a result, a small quantity of borings
from each bar was forwarded to Mr. J. P. Walton, Assoc. R.S.M., of
the Wishaw Iron Works, and his analyses afford strong confirmation
to my own conclusions. In some cases, especially with manganese
and phosphorus, the agreement is not so good as might be desired,
though in each case the general direction of the change is correctly
indicated. But in forwarding his results Mr. Walton said, “I am
sorry that the quantity of material does not allow of my ordinary
processes being followed in every case.” The same remark would
apply to my own analyses; for if the supply of material had been
more abundant the details of the operations would have been modified
in several instances.

The methods adopted in my analyses were as follows. Of each
sample, two quantities, each of 2 grams, were dissolved in aqua
regia, and evaporated to dryness on the water-bath. After heating
at 100° for some time, the residue was extracted with a little hydro-
chloric acid and the silica separated and weighed. To the filtrate
a little barium chloride was added, the solution nearly neutralised,
and allowed to stand at least 24 hours. After filtering off the barium
sulphate, the solution was treated in one case for manganese and in
the other for phosphorus. The first was separated by means of am-
monium carbonate, and precipitated by bromine in concentrated solu-
tion, alkaline with ammonia. The phosphorus was weighed as
pyrophosphate of magnesia after precipitation as molybdate. The
total carbon was determined by combustion, according to the method
I have described (Birmingham Philosophical Society’s Proceedings,
vol. iv, Part II, p. 404; incorrectly abstracted, Abstr., 48, 1161),
while, to save material, the combined carbon was estimated colori-
metricaliy. Of these methods, it should be observed that the phos-
phorus is probably rather low, owing to details in the methods

* At least half of each of the specimens Nos. 8, 12, 14, 15, 16, and 18 have
been returned to Professor Unwin. These pieces may be useful for future reference.

496 TURNER: THE INFLUENCE OF REMELTING

employed, while the values for combined carbon show the direction of
the change rather than the actual amount of this constituent. The
results are given in Table A.

Taste A.—Analyses by T. Turner.

No. of | , —_ : Phos-
melting. Combined.| Silicon. | Sulphur. phorus.

0°40

S616
mo

col]coo|oo|]ooloo!oco
_

wn] ee | ee -
| ~~ Om | mo j—)

2°20

The methods adopted by Mr. Walton for the determination of sul-
phur and silicon were almost identical with my own. The phosphorus
was weighed as molybdate. The results are given in Table B.

Taste B.— Analyses by J. P. Walton.

No. of melting. | Silicon. Sulphur. Manganese. Phosphorus.

ee bot oO
“obo

2orernw

It will be seen that Mr. Walton’s determinations of silicon and
sulphur agree very fairly with my own. In phosphorus, the agree-
ment is not so good, the difference in the eighth and twelfth meltings
being 01 per cent. However, these differences are not such as to

* Varied in parts ; outside white, inside grey.

ON THE PROPERTIES OF CAST IRON. 497

materially affect the iron, and their cause has already been partly
indicated. The most serious divergence is seen in the first melting,
where a difference of 0°5 per cent. of manganese and 0°13 per cent. of
phosphorus occurs. In this case the whole of the specimen was used
in the analysis, so we were unable to check the result in duplicate as
was intended. It must be remembered, however, that owing to the
small quantity of material at our disposal the precipitates of manga-
nese generally weighed but a few milligrams, so that if the results
agree in showing the direction in which the change took place, it is
perhaps as much as can be reasonably anticipated.

From these analyses, I have endeavoured to deduce the probable
composition of the specimens examined. In most cases, the numbers
are the mean of those previously given; but in a very few instances
when any considerable difference was observed, I have ventured to
attach rather more importance to a particular determination which
appeared most trustworthy. The results are given in Table C.

TasLe C. —Composition of Test Bars.

. | | |
No. of | Total | Combined.| Silicon. | Sulphur. | Manga-
| }

melting. carbon. nese.

97 . ‘21 «| O 0°58
98 : 2° 0-23
87 7 0-17
*88 ' "2 0°12

bo bo bo lO bs
eooooo

Or Gr ON
= OD ond a

It will be found on comparing these results with those previously
given by Calvert, that with the solitary exception of the total carbon
in the first melting, our conclusions are entirely at variance. In par-
ticular, it will be seen that the silicon, on the numbers for which my
previous conclusions depended, varies in an exactly opposite direc-
tion, and over a very different range to that which was previously
supposed.

It may be mentioned, however, that doubts have been cast on the
accuracy of Calvert’s analyses by several writers of experience on
this subject, though they had not the opportunity of examining the
specimens. Thus Snelus, writing over 15 years since, in the first
volume of the Journal of the Iron and Steel Institute (p. 37), when
speaking of the results observed in Fairbairn’s experiments, said he
believed the explanation to be, “that at each successive melting the
silicon, and perhaps to a slight extent the carbon, decreases ; but the
iron gradually takes up sulphur and phosphorus from the fuel, and

498 TURNER: THE INFLUENCE OF REMELTING
the deterioration due to these elements more than counterbalances the
increased strength due to diminished carbon and silicon.” In Chemistry
Applied to the Arts and Manufactures, 8, 861, when discussing the
effects of remelting iron in a cupola, the above opinion of Snelus is
put forward, and the remark made, ‘‘ The chemical analyses of some
of these samples, which were made by Calvert, do not appear to sup-
port these conclusions. The analytical results, however, as will be
seen from an examination of the figures, are so remarkable that it is
evident they are not reliable, and it is desirable that they should be
repeated.” It may also be remarked that other analyses published by
Calvert at the same period (Inst. C.H., 1852—53, 354) show that
Eglinton iron was rather highly siliceous, containing 3:12 per cent. of
silicon, while in the present instance it is returned with only 0°77 per
cent.

It will be seen that, like myself, Snelus was of opinion that remelt-
ing had been performed in a cupola. But, in spite of this slight mis-
understanding, the correctness of his explanation is very remarkable.
In my former paper, depending on Calvert’s analyses, I have
endeavoured to account for the effect observed as due to a gradual
increase of silicon, combined with a considerable absorption of sul-
phur in the later meltings. My failure to detect Calvert’s error in
the proportion of silicon was due, in part at least, to a lack of know-
ledge of the influence of definite quantities of each of the other
elements present. This knowledge I have endeavoured to render
more systematic in my last paper, ‘‘On the Constituents of Cast
Iron,” recently read before the Iron and Steel Institute. A com-
parison of the values given in that paper with Calvert’s analysis of
the first. melting, will show that his results are quite untrustworthy,
since his values are those which are characteristic of a white iron,
while the metal operated on was a soft grey.

The action during remelting in a reverberatory furnace is essen-
tially oxidising, and it is therefore natural to expect that the chemical
changes would be the same during remelting as during “ refining,”
puddling, and the open hearth or Bessemer process. This expectation
is realised on examining Table C, the only point of difference being
found in the gradual increase of sulphur.

In Fig. 1 the chemical changes produced during remelting are
represented by curves. These closely correspond in general character
with those given by Windsor Richards, and other authorities, for the
early part of a Bessemer blow. The exception, to which I have
previously referred, noticed in the case of sulphur, is a little obscure
in origin, though the fact is undoubted. The effect of this is of the
utmost importance, and to make it more visible the scale used for
sulphar is 20 times that used for the other elements.

Journ. Chem.S0c.« July I886.

“‘DMOINV) SNILAVW 3S “HAIT SNOS TNOSINUYH

buangjou Jo SMOQUINAT

4 oe + ce 5

ot
ae |

i
= z

a Se
Saal

|
i

él

‘il

FININTIN/

yuan «wad $).t0g

ON THE PROPERTIES OF CAST IRON. 499

If we now glance briefly at the variations of each element and the
effect on the character of the product, we observe—

1. Phosphorus increased from 0°47 to 9°61 per cent. This was
probably due to loss of material during oxidation, the phosphorus
remaining in the metal. A little may also have been gained in other
ways; but I am not aware of evidence to show that an increase of
phosphorus within the limits mentioned would materially affect the
product.

2. Manganese decreased from 1°75 to 0°12 per cent. This would
tend to improve the metal during the earlier meltings.

3. Silicon was reduced from 4°22 to 1°88 per cent. The first effect
of this reduction was to produce softer metal and lower combined
carbon, since silicon was present in quantity in excess of that neces-
sary for the softest metal. On further reducing the silicon, the metal
became stronger and harder. But in these experiments the reduction
was not carried sufficiently far to cause any deterioration due to
deficiency of silicon; for under ordinary circumstances a metal with
1:88 per cent. of silicon would be soft, close-grained, grey, and of very
considerable tenacity.

4. Sulphur increased from 0°03 to 0°20 per cent., and in this
change we have one of the most important alterations which took
place. In the paper recently read before the Iron and Steel Institute,
I have stated ‘‘ that 1 part of sulphur neutralises the effect of at least
from 5 to 10 parts of silicon,” and have suggested that—

For soft foundry iron, sulphur should not exceed ...... 0:13 p. ¢.
, hard or mottled iron, abuut............cceeeeeees 0°20 _=—,,
WHISS ITOM, OVER. 0 cc cccccccccccccccccccccceeseee 0°25 ,,

”

Now in the sixteenth melting we find that silicon, which tends to
produce grey iron, has been reduced to 1°88 per cent., while sulphur,
which tends to produce white iron, has increased to 0°2 per cent., or
just to that amount which is favourable for the production of hard or
mottled iron. The bar itself is described as being white on the out-
side, with about three-sixteenths of an inch grey in the middle, and
therefore agreeing with what might be anticipated from its composi-
tion.

5. The total carbon underwent very slight alteration. The slight
apparent increase at first is quite in accordance with what has been
observed in the puddling process and Bessemer converter.

6. The combined carbon was reduced during the earlier meltings,
but again increased as the metal became harder, reaching a maximum
of over 2 per cent. These changes are dependent on the alterations
in the other elements present, especially sulphur and silicon.

In conclusion, I would express my belief that the experiments of Sir

500 KOHN: SOME AMMONIUM COMPOUNDS AND

W. Fairbairn, with the corrected analyses, afford us valuable informa-
tion as to the influence of chemical composition on the mechanical
value of cast iron. But they do not afford information as to the effect
of remelting, except under the circumstances of these experiments.
This effect will obviously depend on the original composition of the
iron, and the character of the changes which take place in its con-
stituents during remelting. At present I am not aware that we have
any valid grounds for asserting that remelting, apart from chemical
change, is either beneficial or otherwise on the character of the

product.

LII.—Some Ammonium Compounds and other Derivatives of
a-1’ Hydroxyquinoline.

By C. A. Koun, B.Se., Ph.D.

Tue following investigation was conducted under the guidance of
Prof. Otto Fischer, of Erlangen, and is a continuation of his
researches on a-l’ hydroxyquinoline (Bedall and O. Fischer, Ber.
Ber., 14, 442, 1366 ; 16, 712; O. Fischer and Renonuf, ibid., 17, 755),
especially in the direction of the ammonium bases connected there-
with. I take this opportunity of expressing my sincerest thanks for
his invaluable aid and kindly advice accorded to me while engaged in
this work.

The nature of the compounds formed by the action of silver oxide
and the caustic alkalis on the addition products which quinoline,
pyridine, and acridine form with methyl iodide, has long been dis-
cussed, and in spite of the numerous investigations by Claus, Bernth-
sen, Hofmann, and others, still remains to some extent obscure.
These compounds differ in many respects from the quarternary bases
derived from the fatty, and from most members of the aromatic series,
especially as regards their formation by the action of the hydrates of
potassium and sodium on the corresponding ammonium iodides and
in their solubilities in water, ether, benzene, &c.

Whilst the pyridine-ammonium bases possess these abnormal pro-
perties, the classical researches of Hofmann on piperidine and conine
(Ber. Ber., 14, 659, 705) have proved that the reduced pyridines form
true quarternary compounds. This holds good also for the reduced
quinolines, as Feer and Kénigs (Ber. Ber., 18, 2388) have shown to
be the case with hydromethylquinoline (kairolin). Hydroxyhydro-
methyl- and ethyl-quinoline also behave normally as regards their

OTHER DERIVATIVES OF a-1’ HYDROXYQUINOLINE. 501

ammonium compounds except in one point. By the action of solid
caustic potash on the kairin ammonium iodides, an isomeric change
takes place, the added alkyl-group and the hydrogen of the hydroxyl
exchanging places, forming an ether of the original compound.
Hydroxyhydromethylquinoline methiodide when thus treated yields
methoxyhydromethylquinoline. Griess (Ber. Ber., 13, 246, 647)
noticed a similar reaction in the distillation of trimethylphenol-
ammonium, the methyl ether of dimethylamidophenol being formed—

O OCH,
CHC | + CHC ;
N(CHs,)s N(CHs).

While, however, in Griess’s experiments this transference of the
methyl-group occurs in one and the same benzene-ring, the following
results go to show that it can also take place when the hydroxyl] is
situated in an adjoining ring. .

Hydroxyhydromethylquinoline Methiodide, CyH,;NO,CH,I.

Hydroxyhydromethylquinoline, which is obtained by precipitating
the hydrochloride with sodium carbonate, is dissolved in methyl
alcohol, and the solution heated for 1} hours on the water-bath with
a small excess of methyl iodide. The wood-spirit and excess of
methyl iodide are then distilled off, when the ammonium iodide
remains as a crystalline mass. By one recrystallisation from wood-
spirit, with the addition of a little animal charcoal, it is obtained pure
in the form of long white prisms melting at 215—216°. It is easily
soluble in water, sparingly in cold wood-spirit and alcohol, but readily
on warming. On adding ether to an alcoholic or wood-spirit solution
of this compound, it is precipitated in fine white prisms.

Calculated for
Found. C,oH,3N O-CH,L.
Bicsoseesee 41°51 41°64 per cent.

Methoxyhydromethylquinoline.

CH CH,
co Z\°YN cn,

| |

CH Y/0 CH,
C N-CH,
|
U-CH;

The above ammonium iodide distils almost unchanged, a small
quantity of methyl iodide only being given off; but, as has been
VOL. XLIX. 2M

502 KOHN: SOME AMMONIUM COMPOUNDS AND

already mentioned, by the action of caustic potash the iodine is re-
moved and methoxyhydromethylquinoline is formed. The best yield
(about 50 per cent. of the ammonium iodide used) is obtained by
thoroughly mixing 10 grams of hydroxyhydromethylquinoline meth-
iodide with 20 grams of finely powdered caustic potash and heating
the whole in a small retort. After warming for a short time, the
reaction sets in, and the mixture separates into two layers: the upper
oily layer distils over completely on further heating. To purify the
distillate, it is taken up with ether, and the ethereal solution washed
and dried over caustic potash. On distilling off the ether, the new
base is left as a pale yellow oil, which boils without decomposition
at 256—258° (uncorr.).
Calculated for

Found. C,,H,;NO.
© .ncccce S416 74°58 per cent.
|: rere . 848 8°47 ”

The formation of a methyl ether of hydroxyhydromethylquinoline
by the action of caustic potash on hydroxyhydromethylquinoline
methiodide, is to be explained as follows:—The potassium-derivative
of hydroxyhydromethylquinoline methiodide is first formed, and this
decomposes with elimination of potassium iodide, the methyl-group
taking the place of the potassium thus removed, as shown in the
following equations :—

OH IN(CH;), OK IN(CH;).
He H,
+ KOH = + H,0.
/ Ba vr
2 Hy;

OK TN(CH,), | OCH, NCH,

H. A\“~ Fy
Os Qa
/* "haa

9
-

Of the salts of this base, the following deserve mention :—The
platinochloride is obtained in yellow prisms on adding platinum chlo-
ride to a solution of the base in hydrochloric acid. It is readily
soluble in warm water, melts and decomposes at 199°, and contains

no water of crystallisation.
Calculated for
Found. (C,,H,;NO).,H,PtCl,.
Pb. 2000 cose 25°58 25°47 per cent.

By the addition of concentrated sulphuric acid to an alcoholic solu-

OTHER DERIVATIVES OF a-1’ HYDROXYQUINOLINE. 503

tion of the base, an acid sulphate is obtained. It forms white prisms,
which are easily soluble both in water and in alcohol.
Calculated for
Found. (C,,H,;NO),H,S0,.
D ccsvccses - 1181 11°64 per cent.

The hydrochloride is very easily soluble. It is obtained in small
prisms on allowing an alcoholic solution of the base, to which hydro-
chloric acid has been added, to remain for some time over lime and
sulphuric acid in a vacuum.

Methoxyhydromethylquinoline is characterised by the following
reactions :—With sodium nitrite im a slightly acid solution, a carmine-
red coloration is obtained ; on standing, a yellow precipitate forms.
In a dilute sulphuric acid solution of the base, ferric chloride gives a
rose-red coloration, potassium ferrocyanide a white flocculent precipi-
tate.

In order to prove that the product of the action of potash on the
ammonium iodide of methylkairin is really a methyl ether of hydroxy-
hydromethylquinoline, it was prepared directly from the latter. It is
best to start from its sodium salt, which can be obtained by dissolving
the base in soda, and allowing it to stand; after a short time the
sodium salt separates in a crystalline form. ‘This is dried on a porous
plate, and heated with wood-spirit and methyl iodide in a closed tube
for several hours, at 130°. The methoxyhydromethylquinoline thus
obtained was identical with that obtained from the ammonium iodide
by the action of potash. It boiled at 256—258°, gave the charac-
teristic reactions with sodium nitrite, ferric chloride, and potassium
ferrocyanide, and also the platinochloride, which melted with decom-
position at 199°. A little methoxyhydromethylquinoline methiodide
was also formed in the reaction.

Methoryhydromethylquinoline Methiodide, C,,Hi;NO,CH,I.

Methoxyhydromethylquinoline readily combines with methyl iodide
under similar conditions to hydroxyhydromethylquinoline. The
ammonium iodide thus obtained is. easily soluble in water, alcohol,
wood-spirit, and chloroform. It crystallises from wood-spirit in long
white prisms, which melt at 175°, but soften at 173°. When heated
with caustic potash, it distils unchanged (m. p. 175°).

Calculated for
Found. C,,H,,;NO-CH,'I.
ET wccccccees 39°63 39°81 per cent.

KOHN: SOME AMMONIUM COMPOUNDS AND

Methoxyhydrodimethylquinolium Hydrowide.

CH CH,
u ANCA cn,

onl

Y a et

i nw
OCH, CH; OH

By treating the above compound with moist silver oxide, the
iodine is displaced by hydroxyl. The oxide of silver is added till
the solution contains no more iodine ; it is then evaporated down to a
small bulk on the water-bath, at a gentle heat (60—70°) ; and lastly,
further concentrated in the exsiccator. A crystalline mass remains,
consisting of the hydroxide and its carbonate. The hydroxide is
extremely hygroscopic, has an alkaline reaction, liberates ammonia
from its salts, and combines very readily with the carbon dioxide of
the air, so that if the above-mentioned crystalline mass be dissolved in
water and the solution evaporated, well-defined crystals of the car-
bonate will be obtained. The base is very readily soluble in alcohol
and wood-spirit, insoluble in ether, chloroform, and benzene.

With hydrochloric acid and platinum chloride, it forms a double
salt, which crystallises in beautiful yellow prisms. This gave the
following result on analysis :—

Calculated for
(C,;H,;NO,CH,Cl)»,PtCl,.
36°39 per cent.

4°55 a
24°57 .

This salt contains no water of crystallisation, and melts at 200°
with decomposition.

This hydroxide behaves quite normally on heating, being resolved
into methyl alcohol and methoxyhydromethylquinoline. A concen-
trated freshly prepared aqueous solution was heated on an oil-bath, in
a flask. connected with a condenser. Between 100° and 130° an oil,
easily recognised as methoxyhydromethylquinoline, distilled over with
the water, the residue in the flask consisting wholly of this base.
Its boiling pvint was 256—258"; also it gave the platinochloride
(m. p. 199°), and the above mentioned reactions.

The following equation represents this decomposition :—

C,H,(OCH,)N(CH;),-OH = C,H,(OCH,)N-CH, + CH,;OH.

The presence of the methyl alcohol could not be proved with
certainty, owing to its small amount.

OTHER DERIVATIVES OF a-l’ HYDROXYQUINOLINE. 505

Hydroxyhydroethylquinoline Ethiodide, C,,Hi;sNO,C.HsI.

Hydroxyhydroethylquinoline, heated with ethyl iodide in alcoholic
solution on the water-bath, gives this ammonium iodide, which when
crystallised from water forms colourless prisms, melting at 160°. It
is slightly more soluble in alcohol and wood-spirit than the corre-

sponding methyl compound.
Calculated for
C,,H,,NO,C.H,I.

38°14 per cent.

Treated with caustic potash, this compound yields ethoxyhydroethy!-
quinoline, previously obtained by O. Fischer (Ber. Ber., 17, 755), by
the action of ethyl bromide on ethoxyhydroquinoline.

The reaction resembles exactly that with hydroxyhydromethy|-
quinoline methiodide, the ethylkairin distilling over and collecting in
the receiver, where it soon solidifies. Its boiling point was observed
at 269—271° (uncorr.) : it was further identified by the formation of
its characteristic picrate, which crystallises in yellow prisms.

Ethoxyhydroethylquinoline Ethiodide, C,\3HiyNO,C,H;I.

This is prepared similarly to the previous ammonium iodides. It
is very easily soluble in water, alcohol, and wood-spirit. It can be
obtained in long colourless prisms, melting at 136—137°, by adding
ether to its alcoholic solution.

Calculated for
Found. C\3H,9NO,C.H;,I.
35°18 per cent.

Ethoayhydrodiethylquinolium Hydroxide, C\sH,,NO,C,H;-OH.

The above ammonium iodide is readily transformed by means of
moist silver oxide into the corresponding hydroxide, which closely
resembles the methyl compound in its properties. It isequally hygro-
scopic, and absorbs the carbon dioxide from the air with the same
avidity. The platinochloride forms short orange-yellow prisms, which
melt with decomposition at 183°. They contain no water of ecrystalli-

Caleulated for
Found. (CysH,yNO,0,H,Cl).PtCl,
22°23 per cent.

On heating an aqueous solution of the hydroxide, water and a
little ethylkairin distilled over between 100—130°. The residue con-
sisted wholly of ethylkairin. It boiled at 269--271°, and gave the

506 KOHN: SOME AMMONIUM COMPOUNDS AND

characteristic picrate, crystallising in yellow prisms. The distillate
was acidified and redistilled, but no trace of alcohol could be found in
the distillate by means of the iodeform reaction. Hence it is probable
that the alcohol was resolved into ethylene and water, as is known to
be the case on heating tetrethylammonium hydroxide.

Hydroxyhydromethylquinoline Benzylchloride, CjH,;NO,C,;H,Cl.

In order te prove whether, in point of fact, the action of caustic
potash on the methiodide and ethiodide of kairin gives rise to the
transference of the added alkyl-group, I tried to obtain the benzyl
chloride compound of hydroxyhydromethylquinoline. Unfortunately,
however, I was unable to prepare it in a pure state. By heating
hydroxyhydromethylquinoline with benzyl chloride in a closed tube
at 170—180°, a red product is obtained, from which a very minute
quantity of small colourless prisms was separated. The solid product
contained in the tubes was mixed with solid caustic potash and heated,
when a reaction set in similar to that with the methiedide and ethiodide
of kairin, and a yellow oil boiling at about 300° distilled over. This
oil possessed strongly basic properties, and gave a sparingly soluble
amorphous salt with platinum chloride ; this, however, owing to the
small amount of the base available, could not be obtained in a pure
state. Although these results are far from being satisfactory, they
tend to show that the compound thas formed is a benzyl ether of
hydroxyhydromethylquinoline, and therefore that it is the added
group which has been transferred. A methyl ether of hydroxyhydro-
benzylquinoline would not, in all probability, show strong basic pro-
perties. Further, by heating the sodium salt of hydroxyhydromethyl-
quinoline with benzyl chloride, an oil, boiling at 300—325°, was
obtained, ‘which also gave a sparingly soluble amorphous platino-
chloride. The two products, therefore, appear to be identical.

Of the remaining kairin-derivatives to be described, acetylkairin
and hydroxyhydroethylenequinoline possess a physiological as well as
chemical interest.

It is now generally admitted that a quinoline or pyridine group is
directly connected with the constitution of many of the alkaloids, and
this view has been very materially supported by the investigation of
the physiological properties of many synthetically prepared deriva-
tives of quinoline and pyridine.

According to Donath (Ber. Ber., 14, 178—179), quinoline itself pos-
sesses antipyretic, antiseptic, and antizymotic properties, but acts
injariously on the system. The hydroxy-derivatives of quinoline,
prepared by O. Fischer, have been very thoroughly investigated by

OTHER DERIVATIVES OF a-l’ HYDROXYQUINOLINE. 507

Filehne in regard to their physiological action. a-1’ Hydroxy- and
methoxy-quinoline are poisonous, whilst the tetrahydrides of these
compounds, namely, hydroxyhydroquinoline and methoxyhydroqui-
noline, exhibit a physiological action akin to that of quinine, but have
unpleasant secondary actions, such as decomposing albumin, &c.
The activity and instability of these substances were found, as a
result of the above investigations, to be chiefly due to the presence of
the unstable hydrogen-atom united with the nitrogen of the pyridine-
ring, and in order to obtain a more stable molecule, Fischer prepared
his so-called “kairin.” The hydrochlorides of hydroxyhydro-
methylquinoline and -ethylquinoline are capable of reducing the tem-
perature of the body in cases of fever to its normal state, and this
without causing any unpleasant secondary action. Since its dis-
covery (1881), kairin has been used with success as a febrifuge.

a-2’ Hydroxyethylquinoline, a-1’ ethoxyhydromethylquinuoline, and
ulso the hydromethylquinoline (kairolin) of Kénigs, all show a ten-
dency to act like quinine, and differ only in the time the action lasts.
It lasts the longest (15 to 16 hours) with the acid sulphate of a-1’
ethoxyhydroethylquinoline.

The acetyl-derivative of hydroxyhydroethylquinoline is so unstable
that it was useless to investigate its physiological action. The action
of the acid sulphate of hydroxyhydroethylenequinoline was inves-
tigated by Filehne. It is a febrifuge like kairin, but its action is
far weaker, a fact in all probability due to its weaker basic pro-
perties. The investigations on this subject up to the present seem
to show that the relative activity of these febrifuges bears a due pro-
portion to their basicity.

Acethydroxyhydroethylquinoline (Acetylkairin), C,yHyN(O-C,H;0).

Hydroxyhydroethylquinoline is heated with sodium acetate and
acetic anhydride for about three hours, and the whole then poured
gradually into water, the mixture being cooled from time to time.
After long standing, or on the addition of caustic soda solution, the
acetylkairin separates out in the crystalline form. It is insoluble in
water, and is readily purified by dissolving it in cold alcohol and pre-
cipitating by water. It dissolves very readily in ether, crystallising
from this medium in thick prisms, which melt at 63—64°. In pre-
sence of acids or alkalis, it readily undergoes hydrolysis. An alco-
holic solution of acetylkairin decomposes on exposure to the air.

Calculated for
(C,,H,,NO-C,H,0).
71°23 per cent.

7°76 9

508 KOHN : SOME AMMONIUM COMPOUNDS AND

Benzoyl chloride also reacts very readily with kairin forming the
benzoyl ether, a compound no longer possessing basic properties,
easily soluble in ether and alcohol, and very readily decomposed into

its constituents.

Hydroxyhydroethylenequinoline,

CH CH, CH, CH
co 7\°/\ cn, cu, 7 °7N cu

CH L a CH, CH, Ay ) CH

\ ff |
——, Sa

2 mols. of hydroxyhydroquinoline, prepared from hydroxyquino-
line according to the data given by Bedall and O. Fischer (Ber. Ber.,
14, 1366), are heated with 1 mol. of ethylene bromide in a closed tube
at 150° for six hours. The hydrobromide of the new substance separates
in the tube as a crystalline mass, and is best purified by recrystallisa-
tion from wood-spirit. It is thus obtained in small hard crystals,
which are freed from any adhering colouring matter by washing them
with alcohol. To obtain the base, the purified salt is dissolved in wood-
spirit (it is very sparingly soluble both in water and in alcohol), and
sodium carbonate is added, when it separates as a white crystalline
precipitate. It may be readily separated from hydroxyhydroquinoline
by recrystallisation from alcohol, when it is obtained in small silky
prisms. Its melting point was found to be 233°.

Hydroxyhydroethylenequinoline is soluble in alcohol, wood-spirit,
and light petroleum only on boiling, but dissolves in both chloroform
and benzene in the cold. It is insoluble in water and in solutions of
the caustic alkalis. The presence of the bydroxy-groups could only
be proved by warming the base with alcohol till it was partially dis-
solved, and then adding a few drops of caustic soda. The whole
dissolved immediately, the sodium salt separating after a few minutes
in small crystals. These are easily soluble in water, but are decom-
posed almost directly, the unaltered base separating out :—

Calculated for
Found. CoH »4N20>.
74°07 per cent.
741 -

By the linking of the two quinoline-rings, the basic as well as the
acid properties of this compound are materially decreased, Hydroxy-
hydroethylenequinoline dissolves in dilute mineral acids only on
heating; it does not form any salts with acetic, tartaric, or citric acid.

OTHER DERIVATIVES OF a-l’ HYDROXYQUINOLINE. 509

To obtain the hydrochloride, the base is dissolved in chloroform,
an equal volume of alcohol added to the solution, and then strong
hydrochloric acid. After a short time the salt separates in small
colourless prisms, which dissolve easily in wood-spirit, but are only
very slightly soluble in water and alcohol. If platinum chloride be
added to the chloroform alcoholic solution of the hydrochloride, a
platinochloride is formed. This has a faint reddish colour, and is
sparingly soluble both in water and alcohol. An acid sulphate can be
obtained similarly to the hydrochloride; it dissolves readily in water
and crystallises well. This salt decomposes gradually on standing,
with separation of the base; this decomposition, which is ‘readily
effected by heat, is prevented by the presence of dilute acids.

Hydroxyhydroethylenequinoline gives many characteristic reac-
tions. A dilute acid solution gives a yellow precipitate with sodium
nitrite, and a white flocculent precipitate with potassium ferro-
cyanide. By adding ferric chloride to an alcoholic solution of the
base, a beautiful violet coloration is obtained; an excess gives a
brown-red colour. Ferrous sulphate also colours the alcoholic solu-
tion violet, whilst nitric acid produces with the base a carmine-red
colour, which turns yellow on standing.

Dinitroethoryhydroethylquinoline, C\s3Hy,NO(NO,)>.

The nitration of ethylkairin is readily effected, two nitre-groups
being introduced, as is also the case with hydroxyquinoline (Bedall,
Inaug. Dissert., Munich). Ethoxyhydroethylquinoline is dissolved in
strong sulphuric acid, and this solution is added to the requisite quan-
tity of nitre also dissolved in sulphuric acid. The reaction is not at
all violent and the mixture only becomes slightly warm. The product
of the reaction is then poured into water, and the dinitroethylkairin
precipitated with caustic soda. It is obtained in beautiful yellow
prisms by recrystallisation from alcohol, and melts at 76—77°. This
nitro-compound is basic and dissolves readily in strong acids.

Calculated for
Found, C\3H,;NO(NOg)>.
14°24 per cent.

Feer and Kénigs (Ber. Ber., 17, 2388) nitrated hydromethylquino-
line under similar conditions to the above, and obtained a mono-nitro-
compound, so that the presence of the ethoxy-group in ethylkairin is
probably the cause of its great activity.

LIII.—f-Sulphophthalic Acid.
By A. Réz, Ph.D.

Attuoven phthalic acid and several of its derivatives have been
carefully examined, notably nitro-amido- and hydroxy-phthalic acids,
our knowledge of the properties and preparation of the two
possible sulphophthalic acids has hitherto been very limited. Apart
from the interest which attaches to the last-mentioned acids in them-
selves, it seemed probable they would also be suitable for the prepara-
tion of the two hydroxyphthalic acids. Some 20 years ago, Loew
(Annalen, 143, 259) published a short notice on the action of sul-
phuric anhydride on phthalic acid in sealed tubes at 100—105°. He
obtained a small amount of an acid, but its properties did not corre-
spond with those one would expect in a sulpho-conjugated phthalic
acid. According to Loew, on boiling an aqueous solution of the acid
or of its salts, decomposition ensued and sulphuric acid was
liberated.

Jacobsen (Ber., 14, 42) obtained an acid salt of -sulphamine-
phthalic acid by oxidising ortho-xylenesulphonamide with potassic
permanganate.

The same salt, the formula of which Jacobsen gave as—

SO.NH,(4)

Cae} COOK (2)+2H,0.
COOH (1)

was prepared by Remsen and Comstock (Amer. Chem. J., 5, 106)

by the oxidation of naphthalene-8-sulphonamide, whilst naphthalene-

a-sulphonamide gave derivatives of a-sulphophthalic acid.

Miiller and Laiblin (Ber., 18, 1126) were the first to observe that
B-sulphophthalic acid is formed by the action of nitric acid on dinitro-
a-naphtholsulphonic acid, the potassium salt of which is well known
in commerce as naphthol-yellow S., the formula of which, according to
Lauterbach [* Itecherches sur quelques dérivés de Vacide monosul-
fonique du dinitronaphtole.” Geneva, 1882], is CoHy(NO,),(OK)-SO,K.
As they neither published nor intended to make further use of this
interesting discovery, we took up the study of the above reaction
with the view of preparing considerable quantities of pure A-sulpho-
phthalic acid. Only one difficulty was encountered, namely, that
the residue left after oxidation with nitric acid was always yellow.
We cannot account for this, except by assuming that there are traces
of dinitro-2-naphthol (Martius yellow) in the naphthol-yellow S. used.
Naphthol-yellow S. is prepared by tke action of nitric acid on

REE: §-SULPHOPHTHALIC ACID. 511

a-naphtholtrisulphenic acid at a temperature not exceeding 50°.
(German patent No. 10785). Now the latter also contains traces of
naphtholdisulphonic acid, and this with nitric acid forms dinitro-
naphthol, which, in its turn, is only attacked by boiling with con-
centrated nitric acid.

Preparation of B-Sulphophthalic Acid from Naphthol-yellow S.

Naphthol-yellow S. is heated in the water-bath with thrice its
weight of nitric acid (sp. gr. 1‘33—1°38). As soon as the tempera-
ture rises above 50°, a violent reaction sets in, which continues even
without further application of heat, until almost all the colouring
matter is oxidised. The solution is then heated until the evolution
of gas ceases and evaporated to dryness. The yellow residue, con-
sisting of monopotassic A-sulphophthalate and potassic nitrate, is
dissolved in het water, and to this so much of a solution of baric
chloride is added as corresponds with a primary salt of sulphophthalic
acid. After filtering off small quantities of baric sulphate, the solu-
tion is evaporated to crystallisation and the acid barium salt re-
crystallised from hot water (to which a little hydrochloric acid has
been added) until the colour no longer diminishes in intensity. The
baric salt is thus obtained almost white, and is converted into the free
acid by exact precipitation with sulphuric acid. On evaporating the
tiltrate—finally over the water-bath—until syrupy, and then drying
in an air-bath below 100°, the hydrate of B-sulphophthalic acid is
obtained as a crystalline mass of a greyish colour.

If B-sulphophthalic acid is to be used merely for the preparation of
B-hydroxyphthalic acid, it is unnecessary to purify it by means of its
barium salt. As, however, the presence of potassic nitrate would be
disadvantageous in a fusion with caustic soda, it is advisable first to
remove the potassium from naphthol-yellow 8. by heating it with a
considerable amount of hydrochloric acid ; on cooling, dinitronaphthol-
sulphonic acid separates in fine needles, and is then oxidised by
heating with three parts of nitric acid (sp. gr. 133—1°38). On
evaporating, free sulphophthalic acid is obtained, which can be easily
converted into hydroxyphthalic acid, as described in the following
communication (p. 522).

The yield of sulphophthalic acid is quantitative, but a slight loss is
entailed by purifying it by conversion into the baric salt.

Preparation of 8-Sulphophthalic Acid from Phthalic Anhydride.

On repeating Loew’s experiments, it was found that the acid he
obtained was indeed sulphophthalic acid, but contrary to his state-

REE: §8-SULPHOPHTHALIC ACID.

ment, boiling water dissolved the barium salt without decomposition.
Fusion with caustic soda showed that the chief product formed by
the action of sulphuric anhydride on phthalic acid was A-sulpho-
phthalic acid.

As already observed, the publication of Loew on this subject is very
incomplete, and it seemed worth while, therefore, to carry out this
investigation more carefully, especially as phthalic anhydride is easily
obtainable. We were more particularly desirous of improving the
yield of sulphophthalic acid (Loew obtained less than 10 per cent.).

We first experimented in sealed tubes, like Loew, and found that
phthalic anhydride may be almost entirely converted into its sulpho-
conjugated derivatives, by prolonged heating at 170—180° with
fuming sulphuric acid containing a high percentage of anhydride.
Heating under pressure, however, is entirely unnecessary and the
process becomes very simple if carried out as follows :—

100 grams of phthalic anhydride dissolved in 150 grams of sul-
phuric acid containing 20—25 per cent. of anhydride, are heated at
200°, later at 210°, while a constant current of sulphuric anhydride
vapour is passed into the solution. After about 6} hours the reaction
may be looked on as complete, for on diluting a small quantity of the
mixture with water and extracting with ether, not a trace of phthalic
acid can be detected. On cooling, the contents of the flask are poured
into water and the sulphuric acid exactly precipitated with baric
hydrate or carbonate. The filtrate is treated with 5—10 c.c. of con-
centrated hydrochloric acid, then with a hot solution of 100 grams
of baric hydrate, and evaporated to the crystallising point. On
cooling, monobaric 8-sulphophthalate is deposited, and purified by
recrystallising it once from a little boiling water, when beautiful
white crystals are obtained, these being entirely free from a-sulpho-
phthalic acid, an accessory product, which remains in the mother-
liquor.

To separate free §-sulphophthalic acid, a hot aqueous solution of
the baric salt is treated in the way above described under the prepara-
tion of the acid from naphthol-yellow S., and the hydrate of sulpho-
phthalic acid is then obtained as an almost colourless crystalline mass,
representing about 80 per cent. of the theoretical yield, in other words
100 grams of phthalic anhydride gave nearly 150 grams of A-sulpho-
phthalic hydrate.

Before proceeding to a description of this compound, we should like
to make a few observations on a-sulphophthalic acid, which, as
above mentioned, is also produced in this process. The amount being
small, however, escaped notice at first, as the experiments were con-
ducted on a small scale. The presence of another acid besides f-sul-
phophthalic acid was first suggested, when the latter, without being

REE: §-SULPHOPHTH\LIC ACID. 513

first purified by means of the baric salt, was fused with caustic soda.
The melt assumed a bluish-grey tint, this not being the case when
pure f-sulphophthalic acid was similarly treated. After separatirg
the B-hydroxyphthalic acid, which was of course the chief product
formed, we were once successful in obtaining traces of an acid
melting at 155°, and giving an intense bluish-violet coloration with
ferric chloride. Both these facts point to salicylic acid, but they do
not furnish us with sufficient evidence for deciding the question
whether we were really dealing with that acid. The quantity at
cur disposal was too small for further investigation. We, moreover,
never again succeeded in obtaining an acid having the properties of
salicylic acid. It is, however, possible that this acid is formed from
a-hydroxyphthalic acid under favourable circumstances (Ber., 18,
1629), and a-sulphophthalic acid was undoubtedly detected in the
following way :—

The barium was precipitated from the mother-liquors of basic
B-sulphophthalate by a moderate excess of sulphuric acid, and the
liltrate evaporated. After cooling and standing some time, nodules
consisting of small microscopical needles separated—a property never
observed in the case of B-sulphophthalic acid, but one characteristic
of a number of sulphonic acids which are less soluble in very dilute
acid than in pure water. Both Remsen and Comstock (Amer. Chem.
J., 5, 106—111) and Stokes (ibid., 6, 260—262) who worked with
a-sulphephthalic acid, seem to have overlooked this property.

The normal barium salt was analysed. 0°1962 gram dried at 250°
gave 0°1514 gram BaSQ,.

Calculated for
[C,H;(COO).'SO3],Bas. Found.
45°37 per cent.

It was sparingly soluble in hot water, and crystallised in silky needles
which had the appearance of small transparent plates under the
microscope.

The best proof that the above acid is really «-sulphophthalic acid
is, however, adduced from the fact that fusion with caustic soda con-
verts it into Miller’s 2-hydroxyphthalic acid. To this end, a portion
of the mother-liquors from baric £-sulphophthalate were freed from
barium, neutralised with soda, evaporated, and the sodic salts fused
with sodic hydrate at a temperature of 200—210°. The melt having
been softened in water, excess of hydrochloric acid was added, the
solution concentrated and extracted with ether. The residue remain-
ing on evaporating the ether consisted of a- and £-hydroxyphthalic
acids and metahydroxybenzoic acid. The first two could be partly
removed by crystallising the product from a little water; being less

514 REE: S-SULPHOPHTHALIC ACID.

soluble than a-hydroxyphthalic acid, they came out first. The
mother-liquor was then evaporated, the residue dissolved in as little
ether as possible, and precipitated fractionally with benzene, when
8-hydroxyphthalic acid was the first to separate, the solution finally
containing only the a-hydroxy-acid. After distilling off the ether and
benzene and crystallising from water with use of animal charcoal,
short hard prisms were obtained giving all the reactions which Miller
(Annalen, 208, 247) and Jacobsen (Ber., 16, 1962) describe as
characteristic of «-hydroxyphthalic acid. The melting point was
195°, the anhydride melting at 146°.

An analysis of the silver salt gave the following results :-—0°2028
gram yielded 01099 gram AgCl.

Calculated for
C,H;(OH) (COOAg),. Found.

54°54 per cent. 54°19 per cent.

The acid was coloured an intense cherry-red by ferric chloride, ani
was readily soluble in cold water.

The hydrate of B-sulphophthalic acid, CHH;(SO;H (COOH), + H.O
consists of long spikes clustered in rosettes if obtained by prolonged
heating in an air-bath at 95°.

Analyses of the substance dried to a constant weight at 95° gave the
following data :—

I. 0°2048 gram gave 0°2723 gram CO, and 0°0535 gram H,0.

II. 0°2672 - 0°3564 " 0°0800 a
IIT. 0°2394 9» 0°3182 1” 0°0668 -

Found.

Calculated for r >
C,H;(SO,H)(COOH); + H;O. ‘iT. I. =i.
C.... 36°36 per cent. 36°28 36:38 36°25 per cent.
H... 333 ” 289 333 3:10 »

The hydrate is very hygroscopic and moderately soluble in aleohol,
insoluble in ether. On acidifying an aqueous solution with a mineral
acid and concentrating, no crystallisation takes place.

The melting point is about 138—140°. The hydrate obtained by
means of naphthol-yellow S., if slowly heated, melts a little above
100°, losing water at the same time; this is probably due to a
slight impurity. By heating to 140°, 1 mol. H,O, corresponding with
that of the hydrate, is given off.

0°4385 gram lost 0°0301 gram H,O0. Calculated, 6°82 per cent.;
found, 6°87 per cent.

The f-sulphophthalic acid formed is a thick syrup.

REE: §-SULPHOPHTHALIC ACID. 515

I. 0°2212 gram dried at 120°* gave 0°3134 gram CO, and 0°0509
gram H,0.
II. 0°1766 gram dried at 140° gave 0°2500 gram CO, and 0°0432
gram H,0.
Found.
Calculated for poo
C,H;(S0;H)(COOH).. L. II.
C.... 39°02 per cent. 38°65 38°62 per cent.
H... 2°44 " 2°58 2°72 *

The anhydride is produced by prolonged heating at 180°.

04028 gram lost 0°0292 gram H,0. Calculated, 7:32 per cent.;
found, 7°25 per cent.

It is a hard, brown, hygroscopic mass.

Notwithstanding its brown colour, it was only slightly decomposed,
for a solution in water contained no sulphuric acid and was readily
decolorised by animal charcoal.

We preferred to employ sulphophthalic acid prepared from phthalic
acid for the experiments for the preparation of hydroxyphthalic acid.
It may be also added that we found the hydrate more convenient as a
starting part than the acid or its anhydride, seeing that it is the least
hygroscopic of the three, and is most easily obtained pure.

The normal baric salt, [C,H;(SO;)(COO),.].Ba,, is formed by adding
baric hydrate to a solution of the acid until neutral, or by boiling the
latter with baric carbonate. According to Loew, an acid salt is
formed in the latter case. It crystallises in fine silky needles or
small brilliant plates, having the appearance of pointed prisms under
the microscope. The salt contains a varying amount of water of
crystallisation according to the temperature at which it is deposited.
Part of the water of crystallisation is lost on attempting to obtain
the salt air-dry ; it is only entirely removed at 250°.

I. 02042 gram dried at 250° gave 0°7568 gram BaSQ,.
IL.¢ 0°2021 gram gave 0°1568 gram BaSQ,.

Found.
Calculated for aa
[C,H;(SO;) (COO).]} Ba3. I. II.

Ba.... 45°82 per cent. 45°75 45°62 per cent.

Once separated, the normal salt is very sparingly soluble in either
cold or hot water, insoluble in alcohol.

* The product used for Analysis I was obtained from naphthol-yellow S., and
lost its water of crystallisation at 120°.

+ A salt, prepared by boiling the acid solution with baric carbonate, was used for
this analysis,

516 REE: S-SULPHOPHTHALIC ACID.

The dilaric salt, C,H;(SO;ba)(COOba)(COOH)* + 2aq, may be
obtained pure by dissolving the normal salt in water containing so
much hydrochloric acid as is necessary for the formation of the
dibaric compound, evaporating, and recrystallising from hot water.
On cooling, long slender needle-sliaped crystals are deposited.

0°4883 gram of the air-dried salt lost 0°0427 gram H,O on heating

at 150°.
0 2220 gram dried at 150° gave 0°1362 gram BaSQ,.

Calculated for

C,H,(SO ba) (COOba) (COOH) + 2aq. Found.
8°74 per cent. 8°70 per cent.

Calculated for
C;H;(SO ba) (COOba) (COOH). Found.

35°96 per cent. 36°06 per cent.

On dissolving either this salt or the neutral one in a slight excess
of dilute hydrochloric acid and evaporating—

The monobaric salt, CsH;(SOsba) (COOH), + 23H,0, crystallises on
cooling in a characteristic way, the crystals having the appearance of
short pointed prisms grouped in rosettes. The same salt is pro-
duced by evaporating solutions of baric chloride and of the acid; in
the presence of a large excess of hydrochloric acid, however, baric
chloride crystallises out, and not baric sulphophthalate.

The mono-salt is more readily soluble in water than the di-salt,
especially in hot water. Thus 14°76 grams of boiling water when
saturated with the salt contained 7°35 grams of the latter, and 33°6
grams of water at 15° were found to contain only 1°6 gram of the salt
when saturated; whence 1 part of the air-dried salt requires for
solution about 2 parts of boiling water and 21 parts of water at 15°.

The following numbers are the result of two analyses, I being that
of a salt obtained by the oxidation of naphthol-yellow S.; II, that of
one obtained by the action of sulphuric anhydride on phthalic

anhydride.
I. 0°6711 gram of the air-dried salt lost on heating at 150° 0-084
gram H,0. as
2. 0°4516 gram dried at 150° gave 0°1692 gram BaSQ,.
II. 0°4064 gram of the air-dried salt lost on heating to 150° 0-0510

gram H,O.
2. 0°1640 gram dried at 150° gave 0°0618 gram BaSQ,.

* ba = ;
2

REE: §-SULPHOPHTHALIC ACID.

Found.
Calculated for ——
C,H;(SO,ba)(COOH),+24$H,0. I. II.

TTTTT 12°55 per cent. 12°65 12°55 per cent.

Found.
Calculated for

povenlemmran
C,H,(SO;ba) (COOH).. I. II.
Ba...... 21°85 per cent. 22°02 22°15 per cent.

As will be seen from the formula of this salt, it should be possible
to convert it into a compound which is both salt and anhydride, and
this is indeed easily accomplished by heating it at 250°.

0°4549 gram dried at 150° lost, on heating at 250°, 0°0260 gram H,O.

01588 gram dried at 250° gave 0°0626 gram BaSO, Calculated
for loss of 1 mol. of water, 5°71 per cent.; found, 5°74 per cent.

Calculated for
co
C¢H3(SO ba) <co >0. Found.

cocccece 23°18 per cent. 23°18 per cent.

The baric anhydro-salt is a white powder, and is easily converted
by boiling water into the original mono-salt.

The normal potassic salt is very soluble in water. It may be
obtained by adding alcohol to a solution of sulphophthalic acid which
has been neutralised with potassic carbonate. The oil which is first
deposited solidifies after some time. The monopotassic salt,

C,H,(SO;K)(COOH) + 2H,0,

separates in long, brilliant needles on evaporating solutions of sulpho-

It phthalic acid and potassic chloride. It was recrystallised and
ad analysed.
t 1:0410 gram of air-dried salt lost 0°1183 gram H,0 on heating at
rr 150°.
: 0°1736 gram dried at 150° gave 0°0530 gram K,SQ,.
Calculated for

C,H;(SO;K)(COOH), + 2H,O. Found.

- BH i .§ Pirceceees 11-25 per cent. 11:36 per cent.
Calculated for
C;H;(SO3K) (COOH). Found.

— —_s_—sO an eensien 13°73 per cent. 13°69 per cent.

It is readily soluble in water, insoluble in alcohol.
The normal sodic salt is very soluble, and crystallises in micro-
scopical needles. The normal calcic salt, formed by adding calcic
chloride to a solution of sulphophthalic acid previously neutralised

VOL, XLIX. 2N

518 REE: S-SULPHOPHTHALIC ACID.

with ammonia, is an amorphous powder readily soluble in water. The
normal plumbic salt is obtained as an amorphous powder on mixing
aqueous solutions of plumbic acetate and sulphophthalic acid. It is
almost insoluble in water. Hot water containing acetic acid dissolves
it, and on cooling the same salt separates.

The di-ammonic salt, CsH;(SO;NH,)(COO-NH,)-COOH + 13H,0.
On evaporating aqueous solutions of £-sulphophthalic acid which have
been neutralised or made alkaline with ammonia, the solution is
found to have an acid reaction after a short time. To prepare the
pure salt, it is advisable to neutralise two-thirds of a solution of
sulphophthalic acid with ammonia, then to add the other third and
evaporate over the water-bath. After drying a short time at 100°,
the mass is pulverised and boiled with 20 times its weight of
alcohol, to which water is gradually added until the powder has just
dissolved. On cooling, the solution deposits very fine needles, which
are recrystallised in a similar way for analysis.
0°2920 gram of air-dried salt gave 0°0258 gram.
Calculated for
,H;,(SO,;NH,) (COO-NH,)-COOH +1}H;0. Found.
eee 8°80 per cent. 8°84 per cent.

That this is really a di-salt was proved by the following analyses
of the substance dried at 100° :—
I. 0°2010 gram gave 0°2526 gram CO, and 0°0812 gram H,0.
II. 0°2442 gram gave 22 c.c. N at 13°, and 719 mm. bar. pressure.

Calculated for —_—"—
C,H,(SO,NH,)(COO-NH,)-COOH. i. II.
C.... 34°29 per cent. 34°27 — per cent.
H... 429 - 4°48 — =

N... 10:00 10°08

”

It seemed likely that by heating above 190° an imide of #-sulpho-
phthalic acid would be formed. Below 150° there was no loss in
weight, but on further heating, especially between 165—190°, a
decrease corresponding with 2 mols. H,O was observed. We were
unable by gradually raising the temperature to effect the loss of only
1 mol. H,0.
2°6128 grams dried at 100° lost, on heating at 190°, 0°3402 gram
H,O. Calculated for the loss of 2H,O 12°85 per cent., found
13°02 per cent.

The question now arose whether the substance thus formed is
really the ammonic salt of an imide—

KEE: 8-SULPHOPHTHALIC ACID. 519

C,F,(SONH,)<GQ>NH, or CrHy(SO.NH)<G (qq) >O-

or whether it must be regarded as a diamide of the formula—
C.H;(SO.NH,)(CONH;)-COOH.

As its reaction is neutral, however, this precludes the possibility of
the latter alternative, and the compound may therefore be looked on
as ammonic sulphophthalimide. It is purified by crystallising it from
alcohol in the same way as described for ammonic sulphophthalate,
animal charcoal being employed to decolorise it. By evaporating in
a vacuum over sulphuric acid, brilliant monosymmetrical prisms were
obtained.

Analyses of the salt dried at 100° gave the following results :—

I. 02832 gram gave 30°6 cc. N at 16°, and 720 mm. bar.
pressure.

II. 0°1994 gram gave 20°5 e.c. N at 10°, and 712 mm. bar.

pressure.
Found.

Calculated for pote
C,;H3(SO3;NH,) (CO).NH. cB Ti.
11°89 11°53 per cent.

Its aqueous solution is gradually decomposed by boiling, and the
salt is sparingly soluble in boiling alcohol (95 per cent.). Heated in
a bent tube closed at one end, it melts at about 300°, sulphur dioxide
being evolved and the greater part carbonised; at the same time,
small glittering plates collect in the colder part of the tube, con-
sisting, as their melting point (220°) and other properties showed, of
phthalimide.

In order to obtain sulphophthalimide itself, it seemed practicable
to prepare the lead salt. By evaporating aqueous solutions of
ammonic sulphophthalimide and acetate of lead over the water-bath
sulphophthalate of lead was deposited, as an analysis showed (calcu-
lated 56°10 per cent. Pb; found, 55°65 per cent.). To prevent
decomposition, very dilute aqueous solutions (1:5) of the above
products were mixed and allowed to evaporate slowly over sulphuric
acid ina vacuum. Very brilliant fine prisms were obtained in this
way, sparingly soluble in cold water, moderately in warm water.
Unfortunately, however, this product—of most inviting appearance—
proved not to be a salt of sulphophthalimide. It lost 10°13 per cent.
of water at 100°, and the salt dried at 100° was found to contain
42°21 per cent. and 42°04 per cent. of lead.

Professor Soret was so kind as to measure the crystals of ammonic

B-sulphophthalimide. He sent us the following account :—‘ Small
2n2

520 REE: 8-SULPHOPHTHALIC ACID.

white crystals, apparently well formed. The angles, however, differ
considerably in different specimens. The average of several determi-
nations is adjoined.

Monosymmetrical prisms.
ZX = 92° 56’.
a:b:c = 0°70766 : 1 : 0°51903.

“Faces observed: (110) predominating and fairly developed ;
(100) uneven; (001) smaller, but constant ; (111) varying; (001) very
small and rare. The crystals are elongated in the direction of the Z
axis—

Angles of the Normal.
(110) : (110) found 70° 30’* calculated —

(011): (011) ,, 54 49* ‘ —
(110): (011) ,, 72 23* ‘ —
(100): (011) ,, 86 58 . 87° 24’
(100): (111) ,, 54 58 ‘s 55 5
(111): (111) ,, 43 59 ‘ 44 24
(110):(111) ,, 46 56 " 46 39

“The optical axes are in the principal section; through (100) only
a part of the hyperbole can be seen, and this points to an axis which
is considerably inclined to Z (negative).”

Two chlorides of B-sulphophthalic acid were prepared, a trichloride
and a monochloride.

The Trichloride, CsH,(SO,C1)(COC1)».

This is best obtained by heating B-sulphophthalic hydrate (1 mol.)
with phosphorus pentachloride (4 mols.) in a closed tube at 150°.
After distilling off the oxychloride, the trichloride remains as a thick
oil.

0°2467 gram gave 0°3968 gram AgCl.

Calculated for
C,H; ($O,Cl) (COCT)s.

39°52 per cent.

Found.
39°78 per cent.

When treated with cold water, the chloride is gradually converted
into the solid monochloride.

On distilling it, decomposition ensues with considerable carbonisa-
tion, sulphur dioxide is evolved, and a deep yellow oil distils over;
this, however, is not the trichloride of sulphophthalic acid, as after
‘heating it with a solution of caustic soda it is possible to extract by

REE: 8-SULPHOPHTHALIC ACID. 52]

means of ether a substance which, after redistiJlation and recrystal-
lisation from ether, is obtained in brilliant white crystals melting at
6°. On dissolving this in water, evaporating and drying, the melting
point rose to 148°. As these two melting points represent those of a
chlorophthalic anhydride and its corresponding acid, we were justified
in assuming that a chloride of the latter acid had been formed by
distilling sulphophthalic trichloride, as represented by the equation—

C.H,(SO,Cl)(COCI), = C.-H,Cl(COCI), + SO,.

For the results of further investigations bearing on this subject we
beg to refer to our next communication.

The Monochloride, CsH;(SO,C1)(COOH))>.

This is easily prepared by heating sulphophthalic hydrate (1 mol.)
with phosphorus pentachloride (3 mols.). After distilling off the
oxychloride, the residue is mixed with ether, well shaken with water,
and the ethereal solution, after drying over calcic chloride, is allowed
to evaporate. A snowy white powder is then deposited, which, under
the microscope, has the appearance of aggregates of needles grouped
concentrically.

An analysis led to the following results :—

I. 02642 gram gave 0°1384 gram AgCl.
IT. 0°1993 gram gave 0°2646 gram CO, and 0°0387 gram H,0.

Found.
Calculated for oon
C,H,(SO.Cl) (COOH)>. Il.
36°29 per cent. ! — per cent.

7”

13°05 -

The chloride is easily soluble in ether and alcohol, insoluble in
chloroform, benzene, bisulphide of carbon, and light petroleum. When
left in contact with water for a considerable time, no perceptible
action was observed.

B-Sulphophthalic Amide, CH;(SO.NH.) (COOH).

To obtain it, the monochloride is carefully treated with an excess of
ammonia and the solution evaporated to dryness on the water-bath.
After decomposing the ammonic salt by again evaporating with a
slight excess of hydruchloric acid, the powdered residue, consisting
of ammonium chloride and the amide, is extracted with ether, and on
evaporating the latter the pure amide remains,

522 GRAEBE AND REE: SOME COMPOUNDS OBTAINED

02011 gram gave 0°2876 gram CO, and 0°0540 gram H,0.

Calculated for
C,H, (SO,N H,) (COOH)>. Found.

39°18 per cent. 39°00 per cent.
2-98 -

”

The amide is soluble.in water, alcohol, ether, and acetic acid,
insoluble in chloroform, benzene, and light petroleum. On evapo-
rating the aqueous solution, small transparent plates were obtained.
As this substance decomposes on melting, it was not possible to
determine the exact melting point ; it lies between 192° and 202”.

Geneva University.

LIV.—Some Compounds obtained by the Aid of B-Sulphophthalic Acid.
By C. Grarse, Ph.D., and A. Rét, Ph.D.

I. B-Hydroxryphthalic Acid.

By the action of nitric acid on phthalic acid, Miller (Annalen, 208,
224) first obtained two nitrophthalic acids, one of which Baeyer (Ber.,
10, 1079) converted into B-hydroxyphthalic acid by means of the
diazo-compound, and from the other Miller (Annalen, 208, 247)
shortly afterwards prepared a-hydroxyphthalic acid. There is con-
siderable difficulty in separating the two nitrophthalic acids, and in
converting them into hydroxyphthalic acids.

It has been already mentioned in the previous paper that in pre-
paring f-sulphophthalic acid for the purpose of converting it into
B-hydroxyphthalic acid, it is advisable to oxidise dinitronaphthol-
sulphonic acid, and not its potassium salt (naphthol-yellow S.). After
all the nitric acid has been expelled, the yellow residue is neutra-
lised with sodium carbonate, and the sodium salt fused with two and
a-half times its weight of caustic soda. The yield of hydroxy-
phthalic acid obtained was almost quantitative when the following
precautions were employed. A quantity representing not more than
15 grams of sulphophthalic acid was fused for 24 hours, the mass
being continually stirred, and maintained of such a consistence that
it was just possible to stir it without difficulty, this being attained by
the addition of water from time to time. The crucible was heated
in an oil-bath, the temperature being maintained at 175°, and never

BY THE AID OF §-SULPHOPHTHALIC ACID. 523

allowed to rise above 180°. The thermometer was placed in the fused
mass and not in the oil-bath. After softening in water and acidi-
fying strongly with hydrochloric acid, the B-hydroxyphthalic acid
was extracted with ether, the ethereal solution treated with a little
animal charcoal, and the ether evaporated; 83—8°5 grams of
B-hydroxyphthalic acid in an almost pure state was thus obtained.
It was purified by recrystallisation from a little warm water, and the
traces of metahydroxybenzoic acid separated by heating it in a
current of carbon dioxide at a temperature of 200—210°, when pure
8-hydroxyphthalic anhydride sublimed. The results of two analyses
of the acid are subjoined, both having been made with a product, not
purified by means of the anhydride, and only crystallised from a
little water. The acid used for analysis I still contained a little
hydroxybenzoic acid, being taken from that portion which crystal-
lised out first.

I. 0°2164 gram dried at 100° gave 0°4236 gram CO, and 0°0682

gram H,0O,
IT. 0°2034 gram gave 03943 gram and 0°0610 gram H,0.

Found.
Calculated for --- oO
C,;H,@H(COOH),. I, If.

52°74 per cent. 53°38 52°87 per cent.
3°50 3 30 9

”

(Hydroxybenzoic acid requires 60°87 per cent. C and 4°35 per
cent. H.)

The pure acid was found to possess the characteristic properties
described by Baeyer, the melting point being 185°, that of the
anhydride 165°. Baeyer states (Ber., 10, 1079) that 6-hydroxyphthalic
acid is not changed by heating with dilute sulphuric acid at 180°.
Now in the fusion with sodic hydrate, especially at temperatures
above 200°, we found that the acid is partly converted into meta-
hydroxybenzoic acid; also from analogy with f-amidophthalic acid,
it seemed unlikely that 6-hydroxyphthalic acid should be so stable
when heated with dilute acids at 180°: our experiments confirm this
view. 1gram of the hydroxyphthalic acid was heated at 180° for six
hours with 20 grams of hydrochloric acid (1 conct. acid: 1 water). The
tube was opened, and the crystalline mass which had separated from
the acid solution was found to consist of pure metahydroxybenzoic
acid, melting at 200°, having a sweet taste, and giving no coloration
with ferric chloride. To remove all doubt as to the acid being meta-
hydroxybenzoic acid, the ethylic salt was prepared, and as expected,
this crystallised from water in plates, melting at 72°.

This decomposition of B-hydroxyphthalic acid with formation of

524 GRAEBE AND REE: SOME COMPOUNDS OBTAINED

metahydroxybenzoic acid is in accordance with observations made
with several other substitution products of phthalic acid. There
always seems to be a tendency for the production of stable meta-
hydrobenzoic acid, and not of salicylic or parahydroxybenzoic acid.

f-Hydroxyphthalic acid may be obtained from phthalic acid
through the sulpho-derivatives (see preceding communication, p. 513).
After removing excess of. sulphuric acid with baryta or lime, the
a-and f-sulphophthalic acids are converted into their sodium salts and
fused with sodic hydrate as above described.

The proportion of a-sulphophthalic acid in the mixture of the
sulphonic acids is but small. The a-acid is chiefly converted into
a-hydroxyphthalic acid at or above 200°, moreover this acid is more
soluble in water than the corresponding B-hydroxy-acid. No difficulty
is therefore experienced in obtaining the latter pure and free from.
a-hydroxyphthalic acid, the reaction with ferric chloride serving as
a test of its purity; whereas concentrated solutions of 6-hydroxy-
phthalic acid give only a faint yellowish-red tint with that reagent,
traces of the a-acid are readily detected by the deep reddish-violet
coloration produced,

Methylic B-Hydrozyphthalate, CsH;(OH) (COOCHS)>.

This is prepared in the usual way by saturating an alcoholic
solution of the acid with hydrogen chloride, or by the action of methyl
iodide on the silver salt. It crystallises from a hot aqueous solution
in plates melting at 102°, and is soluble in the solvents usually
employed.

I. 0°1751 gram dried at 100° gave 0°3638 gram CO, and 0°0721
gram H,0.
IJ. 0°1674 gram gave 0°3497 gram CO, and 0°694 gram H,0.
Found.

Calculated for aot
C,H;(COOCH,),-OH. 1. Il.

57°14 per cent. 56°66 56°98 per cent.

4°57 4°61 -

”

8-Hydroayphthalimide, CsH3(OH) <CO>NH or
CeH3(OH)< QQ, . NH)??:

a» 9

Dry ammonia gas was passed into fused 8-hydroxyphthalic an-
hydride, which also contained traces of metahydroxybenzoic acid,
until no more gas was absorbed. The cooler part of the vessel was

BY THE AID OF 8-SULPHOPHTHALIC ACID. 525

soon filled with brilliant needles, and towards the end of the operation
a deep yellow oil distilled. It is advisable to stop the current of
ammonia as soon as this yellow oil is observed. The crystalline mass
is dissolved in boiling water, and the solution filtered to remove
carbon and a substance which dissolves in alkali with a deep yellowish-
red colour, but which was not examined further. The solution, on
cooling, deposited small yellowish-white plates or needles, melting at
288—290°. These were sublimed, but the sublimate was still coloured,
and did not become white even after repeatedly recrystallising it from
water.

B-Hydroxyphthalimide is sparingly soluble in cold alcohol or ether,
insoluble in chloroform, benzene, light petroleum, or carbon bi-
sulphide. Hot acetic acid dissolves it readily. It closely resembles
phthalimide.

One of us published a method some time ago (Ber., 17, 2599) for
preparing phthalide from phthalimide, and we were interested in
knowing whether this method might be used as successfully for
the production of B-hydroxyphthalide from f-hydroxyphthalimide :
5 grams of the latter compound and 10 grams of water were heated
over the water-bath with 10 grams of finely divided tin and 50 c.c.
of concentrated hydrochloric acid, the latter being gradually added
until all the tin had dissolved. Strips of zine foil were then immersed
in the solution, and finally boiled with it to precipitate all the tin,
hydrochloric acid being added, if necessary, to keep all the hydroxy-
phthalidine in solution. The nitroso-derivative of the latter was
then prepared by the gradual addition of sodium nitrite, when it
separates as a finely crystalline yellow powder, melting at 170°.
It gives Liebermann’s reaction perfectly, and is only sparingly soluble
in water, but readily in alcohol; almost insoluble in ether.

B-Hydroxyphthalide

is prepared from nitroso-8-hydroxyphthalidine by decomposing it
with caustic soda while still moist, and the yellowish-red solution is
heated gently until nitrogen ceases to be evolved ; excess of hydro-
chloric acid is then added, when the greater part of the B-hydroxy-
phthalide separates. This is recrystallised from hot alcohol, the
solution having been previously boiled with animal charcoal. The
crystalline mass thus obtained consists of pointed prisms clustered
in rosettes. The product is yellow, and the impurity is not removed
by recrystallisation from alcohol or from boiling water.

02159 gram gave 0°5052 gram CO, and 0°0783 gram H,0.

= ES EE nae, UT

ee eee

526 GRAEBE AND REE: SOME COMPOUNDS OBTAINED

Calcu'ated for
C.H,(OK)<CN>0.
64°00 per cent.
eesens 4-00

8-Hydroxyphthalide sublimes in brilliant needles, and separates in
the same form from boiling aqueous solutions. Neither ferric chloride
nor alcoholic solutions of caustic potash produce any coloration. It
melts at about 222°, but begins to soften at 210°.

8-Hydroxyphthalide is sparingly soluble in ether or chloroform,
cold alcohol, and acetic acid. It is almost insoluble in bisulphide of
carbon and benzene, entirely so in light petroleum.

We also tried the action of nitric acid on B-hydroxyphthalic acid,
and publish for the present that when heated gently with thrice its
weight of nitric acid (sp. gr. 1°33) a nitro-product is obtained which
is readily soluble in ether and water, and crystallises from either of
these solvents in almost rectangular flat rhombohedrons. It is easily
converted into an amido-derivative.

Found.
63°81 per cent.
4°03

”

II. 8-Chlorophthalie Acid.

Until quite recently, very little was known of the two monochloro-
phthalic acids. One of them was obtained by Auerbach (Juhres-
berichte tiber die Fortschritte d. Chemie, 1880, 862) by passing a current
of chlorine into a cold solution of phthalic acid in excess of alkali.
The same acid has lately been prepared by Solari (private commani-
cation) by oxidising with nitric acid that dichloronaphthalene which
is formed by heating naphthalene tetrachloride. Guareschi (Ber., 19,
134) obtained the same acid from a different dichloronaphthalene
(m. p. 107°).

The other chlorophthalic acid was first mentioned by Alén (Bull. Soc.
Chim., 36, 434), who prepared it by oxidising e-dichloronaphthalene,
Claus and Dehne (Ber., 15, 321) arriving at the same result inde-
pendently of Alén.

The true constitution of these chlorophthalic acids was first proved
by Kriiger (Ber., 18, 1759), who obtained them from chlororthotoluic
acids of known constitution. (Kriiger, however, gives the melting
point of 8-chlorophthalic acid at 130—134°, whereas it is really 148°.)
Unaware of this latter publication, Claus and Miiller (Ber., 18, 3073)
claimed to have proved the constitution of the two chlorophthalic
acids. One of us, equally unaware of Kriiger’s research, then sent
in a short notice to the Berichte (18, 3359), in which it was shown
that the proof brought forward by Claus and Miiller is insufficient,
whereas the fact that it is possible to obtain a chlorophthalic acid of

BY THE AID OF §-SULPHOPHTHALIC ACID. 527

the melting point 148° (the anhydride melting at 96°) from £-sulpho-
phthalic acid by the aid of phosphorus pentachloride, must be looked
on as conclusive evidence in favour of regarding the last-mentioned
chlorophthalic acid as a 8-derivative.

In describing the trichloride of 8-sulphophthalic acid (see preceding
paper), it was mentioned that this was converted by distillation into
A-chlorophthalic chloride. As a considerable amount is carbonised in
the process, we turned our attention to another method, first used by
Kammerer and Carius in their research on methylsulphobenzoic
acid; this consists in the displacement of the sulphonic group by
chlorine, by the action of phosphorus pentachloride at high tempe-
ratures.

On heating the hydrate of 8-sulphophthalic acid (1 mol.) with
phosphorus pentachloride (5 mois.) at 220°, it is entirely converted
into chlorophthalic chloride, phosphorus oxychloride and thionyl
chloride being the only bye-products, unless an excess of pentachloride
is employed, when chloride of sulphur is also produced.

Chlorophthalic chloride boils at 275—276° (uncorr.) and fumes in the
air. It is saponified with difficulty. To prepare larger quantities of
chlorophthalic acid, the hydrate of 8-sulphophthalic acid* was first
converted into the monochloride by heating it with phosphorus penta-
chloride (3 mols.), and the liquid product was then heated with 2 mols.
more of the pentachloride at 220°. There was no pressure in the
tubes, phosphorus oxychloride and thionyl chloride were distilled off,
and the residue treated with strong caustic potash. After acidifying
with hydrochloric acid, the solution was extracted with ether; the
chlorophthalic acid remaining after evaporation is pure enough for
most purposes, the yield being moreover quantitative. To purify it
further, it was distilled, and the anhydride dissolved in hot water.
We found the melting point to be 148°, that given by Claus and
Miiller. The acid is very soluble in water, less so in water acidified
with mineral acids. It is sparingly soluble in benzene, somewhat
more soluble in chloroform or carbon bisulphide, insoluble in light
petroleum.

* It will be generally found more practicable to use the potassium salt of sul-
phophthalic acid for this purpose instead of the hydrate, of which we happened to
have large quantities. To this end, it is most convenient to oxidise naphthol-
yellow S. with nitric acid (see preceding paper), and then to purify the mono-
potassic salt of 8-sulphophthalic acid by repeated crystallisation from hot water
containing a little hydrochloric acid. The potassium salt when heated at 250°, is

converted into a compound of the formula C,H,(SO;K) <bo>9, which is then

heated at 220° in closed tubes with 3 mols. of phosphorus pentachloride. In this
way, chlorophthalic chloride, phosphorus oxychloride, thiony] chloride, and potas-
situn chloride are formed.

528 GRAEBE AND REE: SOME COMPOUNDS OBTAINED

The anhydride is obtained by distilling chlorophthalic acid. It
solidifies in the receiver to a crystalline mass consisting of brilliant
needles. On recrystallising it from ether, very fine brilliant plates,
melting at 96—97°, are obtained. The anhydride boils 12°5° higher
than phthalic anhydride, that is at 294°5° (thermometer enveloped in
vapour, and pressure = 720 mm.).

An analysis gave the following numbers :—

02452 gram gave 0°1933 gram AgCl.

Calculated for
CO
CoHSCl<G9> 0. Found.
19°45 per cent. 19°45 per cent.

It is readily soluble in alcohol, ether, benzene, chloroform, and
acetic acid, fairly soluble in carbon bisulphide. In light petroleum, it
dissolves only on boiling. We are indebted to Professor Soret for a
measurement of the crystals obtained by crystallisation from ether.
He sent us the following account :—

“The crystals are small and very brittle; the various determina-
tions do not quite agree with one another, owing either to the
extreme narrowness of some of the faces, or to the fact that the

opposite faces are not exactly parallel to each other, this being the
case with most of the crystals. They crystallise in prisms belonging
to the anorthic system.

ZX = 109° 11’: XY = 115° 6’: YZ = 108° 40’.

a:b:c = 0°86625: 1: 1°15725.

“ The crystals are elongated in the direction of X, and flattened in
the direction of Y.

“Faces observed: (001) very much developed; (100) fairly de-
veloped and constant; (010)(110) rather small and varying; (011)
(011) more constant; (013) small and varying; (211)(213) very
small.

Angles of the Normal.

(001) : (100) calculated found 57° 13’#
(100) : (010) 53 43*
(001) : (010) 57 30*
(110) : (100) 54 55*
(001) : (001) 70 57
(001) : (001) 31° 41’ 31 37
(010) : (013) 99 4 99 39

“ Through (001) an optical axis may be observed.”

BY THE AID OF §-SULPHUPHTHALIC ACID.

Methylic B-Chlorophthalate, CsH;C1(COOCHS):.

This was prepared in three ways—(1.) By passing dry hydrogen
chloride into a methyl alcoholic solution of chlorophthalic anhydride.
(2.) By the action of methyl iodide on the silver salt. (3.) By the
action of methyl alcohol on the chloride. In each case, on crystal-
lising from light petroleum, fine needles melting at 37° were obtained,
soluble in the solvents usually employed.

The object in preparing this ethereal salt by three different methods
was, if possible, to find an analogy between chlorophthalic acid and
tetrachlorophthalic acid, the latter yielding different ethereal salts
according to the methods used in its preparation (Ber., 16, 860).

This indeed, as one of us has already pointed out, might be

expected if phthalyl chloride is thus constituted: C,H,< on >0.
2

Why different substances were not obtained when different
methods were used we are at a loss to understand. Perhaps we
may venture to account for this unexpected result by assuming that
unsymmetrical ethereal salts obtained by the aid of the chloride are
unstable, and that under certain conditions they are converted into
ethereal salts having a symmetrical constitution.

Ethylic 8-chlorophthalate was obtained from the anhydride, and
crystallises when exposed to low temperatures (—20°). The boiling
point lies between 300—305°.

os CO —

B-Chlorophthalimide, C.HCl< ~>NH or CHCl< QW) >O-

A current of dry ammonia is passed into fused chlorophthalic anhy-
dride until the gas is no longer absorbed. The crystalline mass ob-
tained is dissolved in boiling water, from which, on cooling, small
brilliant needles or plates separate, melting at 210—211°.

8-Chlorophthalimide is very similar to phthalimide. It sublimes
readily, and is precipitated by acids from its solution in caustic potash.
It is, however, less soluble in boiling water than phthalimide, being,
moreover, but sparingly soluble in ether or cold alcohol, fairly soluble
in warm benzene, chloroform, and acetic acid, insoluble in carbon
bisulphide and light petroleum. Supposing phthalimide to be repre-

sented by the formula C.Hi.< OC NEO it does not appear unlikely

that two B-chlorophthalimides might be produced by the action cf
ammonia on 8-chlorophthalic anhydride, respectively represented by
the formule—

jo Ayrsreaqag

slohedl ty ty

roan Some ecm

530 GRAEBE AND REE: SOME COMPOUNDS OBTAINED

The product obtained, however, seemed to be perfectly homo-
geneous, for the melting point, 210—211°, did not alter on repeated
recrystallisation.

‘ C,H;Cl C.H,(OH
B-Chlorofluorescein, CO< ty >C< OH OHO.

1 gram of chlorophthalic anhydride was heated with 1:2 gram
of resorcin at 200° until the mass became hard. This was then boiled
with water, and the residue dissolved in caustic soda and precipitated
with hydrochloric acid. §-Chlorofluorescein closely resembles fluo-
rescein. It is almost insoluble in water, but dissolves in alcohol or
ether when freshly precipitated. It changes to the crystalline form
when dried and kept, at the same time becoming insoluble in alcohol
or ether. It is readily soluble in acetic acid, almost insoluble in chloro-
form, insoluble in benzene. It dissolves in caustic alkalis or alkaline
carbonates with a deep red coloration, which on diluting exhibits a
magnificent fluorescence, exactly resembling that of a solution of
fluorescein itself.

III. Metachloranthraquinone.

Friedel and Crafts’ excellent method enables us to effect various
syntheses of compounds belonging to the aromatic series. In submit-
ting a mixture of f-chlorophthalic anhydride and benzene to the
action of aluminium chloride, we hoped to obtain metachlorobenzoyl-
benzoic acid, which might serve for the preparation of metachlor-
anthraquinone, this substance not having been prepared hitherto.

Behr and Dorp (Ber., 7, 578) were the first to effect a similar con-
densation. They converted orthobenzoylbenzoic acid into anthra-
quinone, employing phosphoric anhydride as the condensing agent,
whereas Liebermann, who used fuming sulphuric acid, obtained
anthraquinonesulphonic acid. Furthermore, vy. Pechmann (Ber., 12,
2126) obtained orthobromanthraquinone by heating orthobromo-
benzoylbenzoic acid with ordinary sulphuric acid, his method being,
with slight modifications, the one we used for the preparation of

metachloranthraquinone.

Metachlorobenzoylbenzoic Acid, CsHsC1(COC,H;)-COOH.

5 grams of fused phthalic anhydride were dissolved in 50 grams
of benzene, which had been previously dried over sodium, and to the
boiling solution 15 grams of aluminium chloride were gradually
added. When hydrogen chloride ceased to escape, dilute hydro-
chloric acid was carefully added, the benzene poured off while still
hot, distilled, and the residue treated for some time with a warm solu-

BY THE AID OF §-SULPHOPHTHALIC ACID. 531

tion of soda. The filtrate was then acidified with hydrochloric acid,
which precipitates a very greasy substance; this was again twice
dissolved in soda and reprecipitated with hydrochloric acid. After
dissolving the precipitate in hot acetic acid, and adding boiling water
until a slight turbidity appeared, the solution was boiled with animal
charcoal ; on cooling, a yellowish oil and perfectly white needles were
deposited, the former becoming partly crystalline after some time.
To remove the oil, the product was spread on a porous plate and re-
crystallised from benzene: the acid was thus obtained in small
brilliant crystals having the appearance of monosymmetrical prisms
under the microscope.

Metachlorobenzoylbenzoic acid melts at 170°, and is readily soluble in
ether, alcohol, chloroform, and acetic acid, sparingly soluble in carbon
bisulphide, almost insoluble in light petroleum, whilst benzene and
other aromatic hydrocarbons dissolve it only on boiling.

0°3586 gram gave 0°1955 gram AgCl.

Calculated for C,H,Cl(COC,H,)-COOH. Found.
13°63 per cent. 13°49 per cent.

Metachloranthraquinone, OHCI<CO>C.H..

1 gram of chlorobenzoylbenzoic acid and 20 grams of pure sul-
phuric acid were heated for 10 minutes in an oil-bath, the tempera-
ture of which was 160—175°. The mixture was allowed to cool
slowly, and then poured into water, the fine grey powder which sepa-
rated being collected next day, washed well with water, and dried.
It was then recrystallised twice from hot alcohol, when it was obtained
in fine felted needles of a yellowish-grey colour, melting at 204°.

Metachloranthraquinone sublimes without decomposition. It is
readily soluble in hot benzene, sparingly soluble in acetic acid, carbon
bisulphide, or hot alcohol.

0°2783 gram gave 0°1600 gram AgCl.

Calculated for C,,H,Cl. Found.
14°64 per cent. 14°22 per cent.

IV. Trimellitic Acid.

This acid was discovered by Krinos (Ber., 10, 1484), Baeyer (Ann.,
Suppl., '7, 40) shortly afterwards describing its preparation from
mellitic acid. Schreder (Annalen, 172, 94) then obtained it by oxi-
dising colophony with nitric acid, and Hammerschlag (Ber., 11, 88)
prepared the acid from @-alizarincarboxylic acid. In each case
isophthalic was formed as a bye-product.

532 GRAEBE AND REE: §-SULPHOPHTHALIC ACID.

We tried whether Victor Meyer’s method for effecting the ex-
change of a sulphonic group for a carboxyl-group is also applicable to
8-sulphophthalic acid, and we found that trimellitic acid is formed
in this way, together with some isophthalic acid.

One part of monopotassic 8-sulphophthalate, and two parts of
sodic formate are fused for a considerable time at a high temperature,
the mass being continually stirred. During the reaction, sulphur
dioxide and volatile organic sulphur compounds freely escape, and the
mass gradually changes its colour from dull green to brownish-
black. The melt is then softened in water, a slight excess of hydro-
chloric acid added, and the solution extracted with ether. The
residue left on evaporating the ether is dissolved in water, ammonia
added until slightly alkaline, and the solution boiled with animal
charcoal. If necessary, the filtrate is once more neutralised with
ammonia, and then evaporated to dryness with a slight excess of a
solution of baric chloride. The residue is well washed with cold
water, which dissolves the barium of isophthalate, and on decomposing
this with hydrochloric acid and recrystallising the precipitate from a
large quantity of boiling water, small fine needles are obtained, which
melt above 300°, thus proving it to be isophthalic acid.

The residue left after the removal of the: barium isophthalate by
treatment with cold water is decomposed with sulphuric acid, the
solution extracted with ether, the latter evaporated, and the residue
dissolved in water. After neutralising it with ammonia, lead acetate
is added, and the lead salt which is precipitated is decomposed with
sulphuretted hydrogen. On evaporating the filtrate from the lead
sulphide, trimellitic acid crystallises in smal] needles, united to form
rosettes, and melting at 219°; when heated to a higher temperature,
the anhydride distils, and separates in small oily drops, each of which
on cooling solidifies in small concentrically grouped needles, melting
at 158°, very characteristic of this substance.

Only 20—25 per cent. of the theoretical yield of pure trimellitic
acid is obtained, this being chiefly due to the very high temperature
necessary for the fusion, a considerable amount of the acid being
decomposed.

The formation of trimellitic acid, the constitution of which must
undoubtedly be regarded as C,H;(COOH), [(COOH), = 1: 2: 4],
is a further proof, if need be, that the constitution of the sulpho-
phthalic acid employed is C,H COOH),SO,H [(COOH), : SO,H =
1:2:4.

LV.—Mereury Sulphites, and the Constitution of Sulphites.

By Epwarp Divers, M.D., F.R.S., and Tersuxicar Sarmuivzv, M.E.,
Imperial Japanese College of Engineering, Tékyd.

Introduction.

THE compounds described in this paper are, in formation and con-
stitation, not ordinary salts, since they can scarcely be prepared by
the ordinary process of double decomposition—that of bringing in
contact a sulphite with a mercury salt having the mercury in a rela-
tion corresponding to what it is to have in the mercury sulphite. The
study of these compounds should form a not unimportant part of the
chapter in chemistry relating to the constitution of salts; for they
are compounds of the non-saturated oxygenous radicle, SO;, with one
of that group of metals—silver, mercury, and copper-—the members of
which are distinguished by their inability to displace hydrogen, the
relative stability of their compounds with nitrogen, sulphur, and the
halogens, and their property of forming more than one series of salts
in which they are strongly basic.* As such, these salts serve admi-
rably to supply facts for the determination of the particular constitu-
tion of non-saturated oxygenous compounds.

Nothing in the properties of the sulphites of the metals had ever,
we believe, been regarded as evidence that they are not dioxylic salts
until notice was called to the matter by one of us in a paper read to
the Society last year. When, long ago, the organic sulphites were
found to be hemioxylic only, there seemed to be differences enough
between them and mineral sulphites to separate them from the
latter, and consequently they were made into a class called sulpho-
nates, whilst the mineral sulphites were left with the dioxylic con-
stitution attributed to them, of which some of their properties were
supposed to indicate the possession. It soon became apparent, how-
ever, that this separation of the organic from the inorganic sulphites
was not based upon any radical difference of constitution, for sulphite
and sulphonate passed smoothly one into the other in reactions of
double decomposition. The effect of recognising this has not been a
happy one, as, instead of establishing anything, it appears to have
unsettled the views entertained concerning the constitution of oxy-
genous salts, and nourished a belief that there is little that can be

* The intimate relationship of these metals, as thus defined, with the adjacent
iron-platinum metals is obvious. To differentiate them from this group would
hardly be in place here.

VOL, XLIX. 20

534 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

definitely said about it. Davy’s theory, that oxygenous salts are
binary compounds of a metallic radicle with a compound haloid
radicle, having been overthrown in the course of the development of
the chemistry of carbon compounds, there has been an evident un-
willingness to admit the force of facts which seem to require that this
theory in a modified form should be in part reinstated. Witness in
proof of this statement the efforts made to minimise the difference in
basic character between the third atom of hydrogen in phosphorous
acid and the other two. There can hardly be said to be any evidence,
however, against certain oxygenous acids (those which are oxidisable
to other acids) being partly or wholly compound haloid acids; and all
that would seem to be wanted is more evidence from inorganic sources
than has as yet been brought forward in order to gain for non-
saturated oxygenous acids a less reserved admission that they have a
partly haloid or non-oxylic constitution. The admission is one which
seems tous pregnant with consequences, such as we indicate in another
section of this paper; for it is in effect that an acid may have two
kinds of basic hydrogen in consequence of constitution or structure.

Origin of Mercury Sulphites—When it might be supposed that
mercuric sulphite would be formed by direct double decomposition, it
does one of three things, according to circumstances ; it forms with
another basylous radicle a double sulphite such as mercuric sodium
sulphite or mercuric hydrogen sulphite; it forms a basic or oxy-
sulphite with oxide of mercury taken from the nitrate or other mer-
enric salt; or it forms mercuroso-mercuric sulphite by suffering
partial hydrolysis into sulphuric acid and its mercury element.
Again, when mercurous sulphite would be expected there appears,
instead, a salt metameric with it, hypomercuroso-mercuric sulphite.
It is in this manner that the following mercury sulphites are formed
in place of mercuric and mercurous sulphites, both unknown :—

Mercurie sodium sulphite
Mercurie hydrogen sul phite ?*. .

Mercuric oxysulphite Hg< >Hg,0H:.

Mercurie mercurous or merenu- SO.-0 -
rosie sulphite \ Hg<go,.0> He240H:.

Mercurie hypomercurous or nat Hg 800s He OH.
539 2°

SOO

History of the Mercury Sulphites.—Mercury sulphites have already
been examined by Rammelsberg, Péan de Saint-Gilles, Hirzel, Wicke,

pomercuresic sulphite

* Known only in solution.

AND THE CONSTITUTION OF SULPHITES. 535

and A. Vogel. St.-Gilles has done most on the subject. Of the sul-
phites described in this paper, only mercuric sodium sulphite and
mercuric oxysulphite have been previously described—both by St.-
Gilles. The other sulphites of mercury described by these chemists
we make out as having no existence, namely, St.-Gilles’ normal
mercuric sulphite, Wicke’s mercurosic hydrogen sulphite, St.-Gilles’
dimercuric sodium sulphite, and Rammelsberg’s mercurous anhydro-

sulphite.

Normal Mercuric Sulphite wnknown in the Separate State.

According to St.-Gilles, normal mercuric sulphite, Hg”SO;, separates
as a white, curdy precipitate, exceedingly unstable, obtained, but not
in the pure state, when a syrupy solution of mercuric nitrate, free
from excess of nitric acid, is treated with a dilute solution of normal
sodium sulphite; either this salt, or mercuric ovysulphite, (OHg,’SOs)2,
or a mixture of the two, may be obtained in this way according as
the mercury nitrate employed happens to be more or less basic. But,
assuredly, St.-Gilles is wrong.

Mercuric Nitrate and Sodium Sulphite-—When normal mercuric
nitrate, in presence of that very small amount of nitric acid which
seems to be essential to its existence even in concentrated solution, is
treated with a solution of normal sodium sulphite, mercuric ory sulphite
is, from the first, precipitated in a pure or nearly pure condition, and,
as a consequence, the mother-liquor proves to be much more acid than
the original nitrate solution. The accumulation of this free acid
soon prevents the precipitation of more oxysulphite, while much
mercuric nitrate yet remains in solution. The addition of more
sodium sulphite is now followed by momentary effervescence of sul-
phur dioxide, and a partial blackening of the white precipitate. This
blackening effect was observed by St.-Gilles, and he has given, as
conditions of successful preparation of the sulphite and of the oxy-
sulphite, avoidance of excess of the precipitant and the employment
of the nitrate solution in the concentrated state and the sulphite
solution in the dilute state.

But when, as above described, some blackening has taken place, it
soon disappears on agitating the precipitate with the mother-liquor,
and, indeed, the mother-liquor may now be used to furnish more pre-
cipitate which, black at first, soon becomes white, provided too much
sodium sulphite is not added. This white precipitate, however,
consists not of the oxysulphite, but of another salt, mercurosic
sulphite. Now this salt contains mercury in the proportion of only
14 atoms instead of 2 to 1 atom of sulphur, and its admixture with
the oxysulphite in some of St.-Gilles’ preparations may be the reason
of his having found that the composition of the precipitate sometimes

202

536 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

approached that of the normal sulphite. There is another cireum-
stance of which St.-Gilles was unaware, and which must have affected
his results: the oxysulphite contains water, and is efflorescent.

The production of a basic precipitate and an acid mother-liquor by
adding normal sodium sulphite to normal mercuric nitrate, settles the
matter so far as concerns the supposed production of normal mer-
curic sulphite in this way; and a word or two on St.-Gilles’ view as
to the conditions which determine the composition of the precipitate
is better given here than when describing the preparation of the oxy-
sulphite. He referred the degree of basic compusition of the pre-
cipitate to the degree of basic composition of the solution of nitrate
used. This view of the matter depends on the supposition that
nitric acid will dissolve oxide of mercury in excess of the quantity
required to form the normal nitrate. It is known, however, that this
property is not possessed by nitric acid; when nitric acid, sp. gr. 1°4,
is treated with a paste of precipitated mercuric oxide and water, dis-
solution can only be carried to a certain extent before a flocculent
precipitate (which is white in total absence of chloride) of basic
nitrate begins to separate, and the filtered solution contains normal
nitrate with a minute quantity of free acid. It is in fact impossible
by any method to get a basic nitrate solution. St.-Gilles, therefore,
must have got the oxysulphite just as we have done, from normal
nitrate, and not, as he supposed he did, from a basic or oxynitrate.

The fact that the formation of the oxysulphite is accompanied by
that of free acid is also evidence as to the non-existence of mercuric
sulphite. The oxysulphite is indeed insoluble in very dilute nitric
acid, and if dissolved in stronger acid it rapidly undergoes a change
to mercurous sulphate. (See further the section “ Mercuric Oxy-
sulphite,” p. 546.)

The substitution of a solution of mercuric sulphate for the nitrate
is, we may assume, attended with a similar result. Mercuric oxy-
sulphite is not, indeed, precipitated, nor perhaps could it be, because
of its solubility in sulphuric acid not exceedingly dilute; this acid
being necessarily present as the means of keeping the mercuric
sulphate in solution. But the mixed solutions, at first clear, soon
begin to deposit mercurous sulphate, which is just what happens with
a solution of mercuric oxysulphite in dilute sulphuric acid. The sub-
stitution of silver sulphite for sodium sulphite makes no difference ;
the insoluble silver sulphite becomes replaced by the insoluble mer-
curic oxysulphite, and silver nitrate forms in the solution, together
with free acid. (See ‘“‘ Mercuric Oxysulphite,” p. 551.)

Sodium sulphite and mercuric chloride also fail to give mercuric
sulphite, the double sulphite of mercury and sodium being formed in

this case

AND THE CONSTITUTION OF SULPHITES. 537

Mercurosic Sulphite and Mercwric Chloride.—By adding mercurosic
sulphite to excess of solution of mercuric chloride, there is obtained,
besides mercurous chloride, what appears to be a solution of normal
mercuric sulphite in mercuric chloride, if not rather one of chloride-
sulphite, Hg(SO,OHgCl)..* The solution becomes acid also, but this
is accounted for by the continual decomposition of the sulphite which
goes on, whereby it becomes sulphuric acid and mercurosic sulphite
again ; this with the mercuric chloride yields more mercurous chloride
and soluble sulphite, and so on. This is the nearest approach to a
solution of the normal sulphite we know of. The reaction by which
it is produced is—

Heg(SO.0),Hg, + HgCl, = Hg(SO,0).Hg + (HgCl).,
while its decomposition by hydrolysis is expressed by—
3Hg(SO,0).Hg + 20H, = 2S0,H, +2Hg(SO,0).Heg,.

Hypomercurosic sulphite behaves like mercurosic sulphite towards
mercuric chloride.

Silver Sulphiteand Mercuric Chloride.—Silver sulphite is decomposed
by mercuric chloride, yielding ultimately silver chloride, mercurous
chloride, and sulphuric acid :—

Ag,SO, + OH, + 2HgCl, = 2AgCl + (HgCl). + SO,H:.

A change of the mercuric chloride is thus effected in the cold which
can only be accomplished by free sulphurous acid in hot solution.
The primary action of silver sulphite, however, is to form mercuric
sulphite or chloride-sulphite, which then suffers hydrolysis into
sulphuric acid and mercurosic sulphite, from which point the decom-
position proceeds as described in the section on ‘* Mercurosic Sulphite.”
By pouring excess of mercuric chloride solution on to silver sulphite
and quickly filtering, a solution is obtained which continues to
deposit mercurous chloride for some time. The primary reaction is
therefore—

2Ag,S0, + 3HgCl, = 4AgCl + Hg(SO.0HgC)».

Sulphurous Acid and Mercuric Oxide—Precipitated mercuric oxide
can be dissolved in excess of an aqueous solution of sulphur dioxide, and
can also under suitable conditions form with it an insoluble compound,
mercurosic sulphite, but not normal mercuric sulphite. St.-Gilles
states that when mercuric oxide in water is treated with sulphur
dioxide, and the white substance produced is very quickly filtered off,

* Mercuric chloride certainly does help to keep mercuric hydrogen sulphite in
solution (see “ Mercuric Chloride and Sulphurous Acid,” p. 557, in section on
Mercurie Hydrogen Sulphite Solution.)

538 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

the filtrate almost immediately deposits a white precipitate which is a
mixture of normal mercuric sulphite with mercurous sulphate. The
precipitate obtained in this way, however, is simply mercurosic
sulphite, like the white substance first produced, and is only liable to
be mixed with a little mercurous sulphate when the filtrate from
which it has deposited has been allowed to stand for a much longer
time than that indicated by St.-Gilles.

Mercurie sodium sulphite, when treated with a little dilute nitric acid
or sulphuric acid, yields an insoluble sulphite of mercury; this,
however, is mercurosic sulphite. (See the next section.)

The only conclusion possible therefore is that mercuric sulphite has
no stability, except in combination with other sulphites, or with
mercuric oxide or chloride.

Mercuric Sodium Sulphite.

Our purpose has not been to include in the investigation which is
the subject of this paper the double sulphites of mercury with other
metals, but it would have been impossible to gain a comprehensive
notion of the relations of the single sulphites, without studying the
properties of at least one of the alkali double sulphites, especially as
normal mercuric sulphite is unknown. Double sulphites of mercury
are readily obtainable, and several of them are known. They are ail
mercuric salts, while the double sulphites of copper are equally all
cuprous.

Mercuric sodium sulphite has already been described by St.-Gilles.
Besides the ordinary sulphite, which is half of sodium and half of
mercury, he has described another which is two-thirds a sulphite of
mercury. The former is the subject of this section of our paper; the
latter is noticed in the short section following.

Properties.—Mercuric sodium sulphite is a comparatively stable
salt, colourless, crystalline, very sparingly soluble in water and
neutral to litmus (St.-Gilles*). It contains water, and its composi-
tion, as ascertained by St.-Gilles, is expressed by the formula
Hg(SO;Na),,0H,. It requires about 25 parts of cold water to dissolve
it. It may be preserved for a time but little changed. It slowly
turns grey, however, and decomposes into mercury, sodium sulphate,
and sulphur dioxide. Heated moderately, in the dry state, it evolves
much sulphur dioxide together with the water it contains; at the
same time it turns brownish and greyish. When the residue is

* We write this section, like the rest of the paper, all from our own experience,
except when we state otherwise, but whenever our observations are not the first
recorded, we udd a reference to the previous observer.

AND THE CONSTITUTION OF SULPHITES. 539

wetted, it blackens, through the formation of a flocculent, brownish-
black matter, in small quantity, the principal products being metallic
mercury and sodium sulphate. The brown-black matter heated with
water changes into a pure black substance, insoluble in hot dilute
nitric acid, and apparently sulphide. The main change effected by
dry heat is probably the same as that by hot water, but a little
mercuric sodium thiosulphate would seem to be also produced, by a
reaction between the sulphur dioxide and some as yet undecomposed
mercuric sodium sulphite. (Comp. Divers, Trans., 1885, 208.)

Heated in aqueous solution, it is converted into mercury, sulphur
dioxide, and sodium sulphate :—

Hg(SO;Na), = Hg + SO, + SO,Na.

The change is a sudden one (St.-Gilles).

According to St.-Gilles, mercurous sulphate is one of the first
products of the decomposition of mercuric sodium sulphite by hot
water, but blackens rapidly and then dissolves, leaving only metallic
mercury. ‘These are his equations modernised :—

2Hg(SO,Na), = Hg.SO, + Na.SO, + Na,8,.0;

We have not observed the formation of mercurous sulphate, and do
not consider it possible that it could be formed. When the mercury
comes out, it has a peculiar, we might say characteristic, appearance,
familiar probably to many chemists; it appears as a bright rolling
cloud, suggesting a precipitate of great volume, an effect due
apparently to metallic reflection of light; this cloud then melts away,
leaving a relatively minute, very dark grey deposit of mercury at the
bottom of the vessel, the bright mist becomes, so to speak, a rain of
mercury. These phenomena must have deceived St.-Gilles. Mer-
curous sulphate would not come out of solution from an abundance of
hot acid liquid, and it cannot exist in presence of sodium sulphite, to
the action of which it is exceedingly sensitive, and, therefore, would
hardly be produced along with it. A single mercury sulphite when
heated with water (mercuric oxide or other mercury compound being
absent), becomes mercury and sulphuric acid; in the double sulphite,
the sulphuric acid produced acts on the sodium sulphite, giving
sodium sulphate and sulphurous acid.

The effect mentioned below of sodium hydroxide, sodium sul-
phite, or sodium chloride, in preventing decomposition of mercuric
sodium sulphite in a hot solution, seems to show beyond doubt that
the decomposition which occurs in absence of all such substances
depends on the presence of acid, developed in minute quantity by the
iucipient decomposition of the salt. As to the sodium chloride, that

540 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

will be effective through its displacing the sulphuric acid from the
sulphite by hydrochloric acid, which is incapable of causing any
liberation of mercury from the double sulphite. Since acid is the
cause of the change, the formation of mercurosic sulphite will precede
that of free mercury. (See below.)

Sodium hydroxide is without action when in dilute solution (St.-
Gilles) ; in its presence a solution of mercuric sodium sulphite can
be boiled without decomposing. If, however, the sodium hydroxide
is added in somewhat large quantity and in concentrated solution, it
precipitates some of the mercuric oxide, slowly in the cold, rapidly in
the hot solution. Some only of the oxide is precipitated, and this is
the scarlet-red form, instead of the usual orange-yellow one. In thus
yielding the red oxide, mercuric sodium sulphite resembles black
mercuric oxychloride. Potassium hydroxide precipitates mercuric
potassium sulphite, which is much less soluble than the sodium salt.
Mercuric oxide is not precipitated from the potassium salt by the most
concentrated solution of potassium hydroxide, even when hot.

Hydrochloric acid changes it into mercuric sodium chloride and
sulphurous acid. Nitric acid (or sulphuric acid) sufficiently diluted,
and not in great excess, yields mercurosic sulphite, sodium sulphate
and nitrate, and sulphur dioxide, thus—

3Hg(SO;Na), + 4HNO, — 4NaNO, + Na,SO, + Hg;(SOs)2
+ 380, + 20H.

In this complex change, the mercuric sulphite will not be decom-
posed by the nitric acid itself, for dilute nitric acid does not decompose
mercury sulphites. The double salt will first exchange its sodium for
the hydrogen of the nitric acid, and then the mercuric hydrogen
sulphite will suffer partial hydrolysis and be converted into mer-
curosic sulphite, sulphuric acid, and sulphurous acid, in the way
described in the section on “ Mercurie Hydrogen Sulphite Solution ”
(p. 556). The addition of alittle acid to a solution of mercuric sodium
sulphite makes it smell of sulphur dioxide, but does not cause
immediate precipitation. Some greyish mercurosic sulphite may soon
precipitate, but in any case mercury is gradually set free. The
acidified solution behaves, in fact, as one of mercuric hydrogen sul-
phite (which see for particulars). Sulphurous acid is without ap-
parent action.

Sudium sulphite, in dilute solution, is without action, but its
presence stops the decomposition of mercuric sodium sulphite by heat ;
in concentrated solution, it precipitates much of the mercuric sodium
sulphite, which is much less soluble in it than in water, slightly
soluble only as it is in the latter. Sodiwm chloride solution is without
action; it prevents (St.-Gilles), or almost prevents, the usual decom-

AND THE CONSTITUTION OF SULPHITES. 541

position when mercuric sodium sulphite is heated alone with water.
The evaporated solution deposits crystals of the mercuric sodium
sulphite unchanged, according to St.-Gilles. Sodium hydroxide has
no effect on the boiling mixed solutions, indicating that there is no
mercuric chloride present in the mixture. Potassiwm iodide is without
action (St.-Gilles), when the sulphite is in solution. Solution of the
iodide added to the solid sulphite gives a little crimson (not yellow or
scarlet) precipitate, not further examined. The mother-liquor, or a
mixed solution of the sulphite and iodide, gives a precipitate of
ordinary mercuric iodide, when treated with sulphur dioxide (see
p- 544).

Mercurie chloride solution, in the cold, is without action; but
when the solutions are hot, mercurous chloride, sodium sulphate, and
sulphur dioxide are produced :—

Cl,Hg + Hg(SO;Na), —- (ClHg), + SO,Na, + SO..
Or, as St.-Gilles found, hydrochloric acid may be formed (and then
there will also be sulphuric acid) —

20H, + 3Cl,Hg + Hg(SO,Na), = 2HCl + 2(ClHg), +
SO,Na, a SO,H,.

This occurs only when the mercuric chloride is in excess, the two acids

then resulting from the well-known reaction between mercuric
chloride and sulphurous acid. (St.-Gilles represents the change
differently.) It is almost certain that the formation of mercurous
chloride is due to the independent decomposition of the mercuric
sodium sulphite, for when this is prevented by the addition of sodium
chloride no change of mercuric chloride to mercurous chloride
takes place. Mercurosic sulphite is formed from the mercuric sodium
sulphite by the action of traces of sulphuric acid, and the mercuric
chloride then acts on it, as described in the section on “ Mercurosic
Sulphite,” to yield mercurous chloride and sulphuric acid again; so
that once started, the decomposition of the sulphite proceeds very
quickly. Mercuric iodide is without action in the cold. Heated, it is
converted into the yellow iodide, and ultimately to mercurous iodide
apparently. The reaction will be similar to that between the sulphite
and mercuric chloride.

Mercuric oxide (precipitated) is without action in the cold, but
warmed with the solution it gradually gives place to somewhat grey
mercurosic sulphite, sodium sulphate being also formed. Most
probably some decomposition of the mercuric sodium sulphite first
occurs, as usual, into sodium sulphate, mercurosic sulphite (or
mercury), and sulphurous acid, and then the sulphurous acid reacts
with the mercuric oxide to form more mercurosic sulphite and sul-

We am

Eee ee ee

42 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

phuric acid, which with sodium sulphite gives sodium sulphate and
sulphurous acid, and so on.

Mercuric nitrate solution (aqueous nitric acid saturated with pre-
cipitated mercuric oxide), in excess, dissolves mercuric sodium sul-
phite, and then very rapidly yields a precipitate of mercuric oxysul-
phite. Provided the mercuric nitrate solution is somewhat concen-
trated (10 per cent. or stronger), and free from avoidable excess of
acid, oxysulphite alone is precipitated if the mercuric sodium
sulphite is added at once in quantity small enough to leave much of
the nitrate undecomposed, and consequently not to generate too much
nitric acid in this solution. The formation of the oxysulphite appears
to be the only reaction of the mercuric nitrate itself, but other
changes occur, owing to the nitric acid set at liberty :—

(OHg,SO,).,*

The mother-liquor of the oxysulphite, treated with more mercuric
sodium sulphite, yields at first mercurosic sulphite, and then
gradually mercurous sulphate. When the quantity of the sulphite
first added to the mercuric nitrate is too great, mercurosic sulphite
may or may not be precipitated, but the oxysulphite which is formed
remains in solution, soon, however, to pass, with marked rise of tem-
perature, into mercurous sulphate, which then crystallises out. In
this case, it is sometimes quite practicable to pour the mother-liquor
off from the mercurosic sulphite, before the mercurous sulphate
begins to separate. Mercurosic sulphite can always be obtained by
adding the mercuric nitrate gradually to the mercuric sodium sulphite.
It is no doubt formed, in either case, through the action of the nitric
acid present on the mercuric sodium sulphite, and of liberated sul-
phurous acid on the mercuric nitrate ; whilst the mercurous sulphate
is the result of a metameric change of the mercuric oxysulphite in
solution in the nitric acid.

Mercurous nitrate solution precipitates mercurosic sulphite :—

Hg(SO;Na), + (HgNO;), = 2NaNO, + Hg(SO;).Hg».

Mercurosic sulphite prepared in this way is always of a light baff
colour, although when formed by hydrolysis it is brilliantly white.
When the nitrate solution is free from nitrous acid, and, so far as
possible, from nitric acid also, only traces of sulphuric acid are formed.
Nitric acid, in precipitating mercurosic sulphite, generates sulphuric
acid, as already explained.

* In the section on “ Mercuric Oxysulphite,” the nature of this remarkable re-
action, and others related to it, is discussed under the heading “ Formation.”

AND THE CONSTITUTION OF SULPHITES. 543

Mercuric sulphate solution does not precipitate mercuric oxysul-
phite, in consequence, apparently, of the solubility of the latter in
sulphuric acid. The mixed solution soon deposits mercurous sulphate.
By adding quickly about an equivalent quantity of the mercuric
sodium sulphite, mercurosic sulphite can be precipitated. By adding
a little strongly alkaline solution of mercuric sodium sulphite to excess
of concentrated mercuric sulphate solution, mercuric oxysulphite
can be precipitated. The mercuric sulphate solution used to prove
this fact was made by saturating rather concentrated aqueous sulphuric
acid with precipitated mercuric oxide, and contained only a very
small quantity of free sulphuric acid.

Mercurous sulphate behaves like the nitrate, when it is shaken in
the solid state with solution of mercuric sodium sulphite. It is then
quickly changed to mercurosic sulphite.

Formation and Preparation.—Mercuric sodium sulphite is
formed in reactions between—

(a.) A solution of sodiwm sulphite, or sodium pyrosulphite (‘‘acid

sulphite”) and any mercury sulphite.

(b.) A solution of silver sodium sulphite and mercuric oxysulphite.

(c.) A solution of sodium hydroxide and any mercury sulphite.

(d.) A solution of sodium chloride and any mercury sulphite.

(e.) A solution of sodium sulphite or pyrosulphite and an owide, or a
halvid salt, or an oxylic salt, whether mercuric or mercurous.
Mercury cyanide is, however, an exception, being here in-
active, although it reacts with ammonium sulphite, according
to St.-Gilles. Yellow and green mercury iodides are also
inactive among haloid salts.

Concerning a, b, c, and d, consult the accounts of the several sul-

phites.

e. Sodium sulphite in dissolving mercuric oxide sets free sodium
hydroxide. Sodium pyrosulphite and mercuric oxide yield no bye-
product. Hirzel first employed mercuric oxide and a pyrosulphite
for preparing mercuric ammonium sulphite. Sodium pyrosulphite
and mercuric oxide afford an excellent means of preparing the double
sulphite in quantity. As much heat is evolved in the reaction, it is
necessary to cool, especially if the solution of sodium pyrosulphite
is concentrated ; if allowed to get hot, the mercuric sulphite becomes
hydrolysed, yielding mercury and sulphuric acid. If excess of
mercury oxide has been added, it can be dissolved by adding a little
more sodium pyrosulphite. The crystalline precipitate of double
sulphite can be washed with a little water and well drained. Should
it contain any mercury, through the solution having been allowed to
get hot, the salt can be dissolved in hot water, to which a little sodium

544 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

hydroxide has been added, and the solution filtered and set aside to
crystallise, or precipitated with alcohol. When sodium sulphite is used
in place of the pyrosulphite, the double sulphite is not decomposed if
the solution becomes hot, and may therefore be at once drained, and
then washed with alcohol, in order to remove sodium hydroxide.

Mercuric chloride is a very convenient source of the double sul-
phite; concentrated solution of the sulphite or pyrosulphite being
poured on to the finely-powdered chloride. As it is exceedingly
soluble in these solutions, no precipitation of double su] phite occurs
until the sodium salt has been added in sufficient quantity. At first
the salts formed are the two double salts (St.-Gilles) :—

2HgCl, + 2Na,SO, = Hg(Cl,Na). + Hg(SO,Na).,

but the further addition of sodium sulphite converts the double chlo-
ride to sodium chloride only, and brings down most of the mercury
as double sulphite. If the mixture is heated when the mercury chloride
is in excess, decomposition ensues (see “Mercurie chloride and
mercuric sodium sulphite,” p. 541). Sodium sulphite at once dissolves
mercuric iodide as it does the chloride, but pyrosulphite has no action
on it, strange to say, whilst sulphurous acid even precipitates mercuric
iodide from its solution in sodium sulphite.* Mercurous chloride and
mercurous oxide behave like the mercuric compounds, but half the
mercury remains in the free state, the other half forming the double
mercuric sulphite. .

St.-Gilles found that mercuric oxide, chloride, and iodide, and
mercurous chloride yielded the double sulphite. But, besides these
haloid salts, oxylic mercury salts, nitrates, and sulphates, both mer-
curic and mercurous, also furnish this compound; they should be
added, mercuric salts especially, to the sodium salt gradually and not
in excess, because in excess they destroy the double salt in the manner
already described. This decomposition led St.-Gilles to believe that
only haloid salts could be used for making the double sulphites. In
the case of the mercurous salts, half the mercury becomes free neces-
sarily, but even with mercuric salts, a very little mereury, at least, is
also set free; it is possible to get even half the mercury of the
mercuric salt free, if the sulphite is not added in sufficient quantity
for some time. The first effect of adding sodium sulphite to mer-
curic nitrate, for example, is to form mercuric oxysulphite and
nitric acid :—

4Hg(NO;), + 2Na,SO, + 20H, = 4NaNO; + 4HNO, + (OHg.S0;),,

* Sulphurous acid also precipitates mercuric iodide from a solution of mercuric
sodium sulphite to which potassium iodide has been added not in excess. It does
not decompose mercuric sodium iodide.

o,n Ss ++ = DDH »

ga M7

AND THE CONSTITUTION OF SULPHITES. 545

and this oxysulphite dissolved in the nitric acid, or even lying in it,
gradually changes to mercurous sulphate; on now adding more
sodium sulphite, this mercurous sulphate yields ultimately only half its
mercury as double sulphite, and half in the metallic state. By using
the sodium sulphite at once in some excess, this is avoided, and almost
all the mercury is converted into the double salt. The nitrates and
sulphates, however, cannot be recommended for the preparation of the
double salt, the oxide and chloride are far more advantageous. The
formation of mercuric sodium sulphite from mercurous salts—nitrate,
sulphate, chloride—is preceded by that of hypomercurosic sulphite,
which is then resolved by more sodium sulphite into mercuric sodium
sulphite and metallic mercury (see “‘ Hypomercurosic Sulphite ”).

Chemical Composition.—We have not considered it necessary
to confirm St.-Gilles’ analysis of mercuric sodium sulphite ; but some-
thing may be here said as to the constitution ascribed to it in this
paper. If instead of Ne<ooree mercuric sodium sulphite had the

formula Se<one that is, had its mercuric radicle been half
2

oxylic, or if it had the formula Hg< oom in which the mercury
is represented as wholly oxylic, the cause of the relative stability of
the double salts, of the non-existence of simple mercuric sulphite,
and of many of the special reactions of the double salts, would cease
to be apparent. There would also be no explanation why mercuric
sodium sulphite has no action on litmus, and is not oxidisable by the
air, whilst sodium sulphite is strongly alkaline and very oxidisable.
With the constitution expressed by the formule Hg(SO,ONa), and
Na(SO,ONa), sodium sulphite is alkaline, because of the instability of
the non-oxylic union of sodium with sulphuryl, in presence of water,
whilst in the mercuric sodium sulphite there is, through the absence
of sodium so situated, the same indifference to litmus as is shown by
sodium sulphate. Sodium sulphite is oxidisable, and mercuric sodium
sulphite is not, because the non-oxylic sodium is oxidisable, and
mercury is not; Na(SO;Na) oxidises to NaOQ(SO;Na), but Hg(SO,;Na),
undergoes no such change.

St.-Gilles’ Dimercurice Sodiwm Sulphite believed not to Exist.

St.-Gilles obtains the mercuric sodium sulphite just described by
adding mercuric chloride to excess of sodium sulphite, and then
evaporating to crystallisation; but, as pointed out in the preceding
section, by using solid mercuric chloride in powder, in place of its

546 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

very dilute cold solution, the double sulphite can be obtained, without
evaporation, as a precipitate.

Besides this double sulphite, he gets another, formulated as
Hg,Na,(SO;);,0H2, by employing hot saturated solutions, the mercuric
chloride being in excess; some mercurous chloride is formed, which
he directs should be filtered off, and the filtrate left tocool. We have,
however, failed to confirm his statements. With solutions hot enough,
a most abundant formation of calomel occurs, and on cooling nothing
separates, except sometimes crystals of the excess of mercuric chlo-
ride. With solutions less hot, the precipitation of calomel is slow
and imperfect, so that the warm filtrate continues to deposit calomel
in crystals, but no new double sulphite.

St.-Gilles asserts that the solution of this double salt has an alka-
line reaction. It is difficult tosee why this should beso. The filtrate
from the calomel which should yield this salt is markedly acid, as a
necessary consequence of the precipitation of calomel having occurred
in it. Indeed, St.-Gilles himself mentions in another place this forma-
tion of acid along with calomel (see the action of mercuric chloride
on the double sulphite in preceding section, p. 541). For our own
part, we can hardly conceive of any way in which such a sulphite
could be formed, considering that mercuric sulphite in the separate
state is unknown.

Mercurie Oxysulphite.

Mercuric oxysulphite was first obtained in 1852, by St.-Gilles, and
is the only one out of the three sulphites of mercury, which have been
described, the existence of which we admit.

Properties.—Mercuric oxysulphite is a curdy, or granular, dense
salt, insoluble in water, and very unstable (St.-Gilles). It has a
faint yellow colour, and its composition is expressed by the formula
(OHg.SO;),OH,. Left to itself in the air at common temperatures,
it changes, in some hours, into mercurous sulphate; but in a vacuum
over sulphuric acid, it can be preserved for days partly undecom-
posed. It is thus rendered anhydrous, but in this state consists
largely of mercurous sulphate: St.-Gilles represents the salt as
anhydrous when precipitated. As pointed out by him, mercuric
oxysulphite (in its anhydrous state) is metameric with mercurous
sulphate. In ordinary moist air, it appears to change completely into
this salt, but when heated even very gradually the change is not so
simple :—

Mercuric oxysulphite. Mercurous sulphate.

SO,OHgO. 7 _ 080.0
Hg<go0,0Hg0> Hs = He<o.g0;.0>Hs-

AND THE CONSTITUTION OF SULPHITES. 547

Mercuric Oxysulphite is Explosive.—This salt is remarkable in being
exceedingly, though not violently, explosive. The explosion is attended
by a moist, mercurial fume, the scattering of powdery particles, and
by a dull sound when in a confined space. A temperature of 73° is
sufficient to cause its explosion, but any quantity of it can at once be
exploded at common temperatures, by touching the least portion of
the whole with a heated body. It can also be exploded on the anvil,
although its decomposition is so gentle that any noise produced is lost
in that of the blow of the hammer. According to St.-Gilles, it is
simply transformed into mercurous sulphate when heated ; this salt is,
indeed, the main product, but mercury and mercuric sulphate are also
produced in quantity, besides water, of course, and a trace of mercuric
oxysulphate. Only when slowly effected by warmth and moisture, is
the change into mercurous sulphate perfect, or nearly so. It seems
best to regard the decomposition of mercuric oxysulphite as being
primarily, like that of the other mercury sulphites, one of hydrolysis.
The first product will then be mercury, sulphuric acid, and mercuric
oxide :—

(OHg,S0,), + 20H, = 20Hg + 2Hg + 2S0,H,.

The acid will then react with the mercuric oxide to form mercuric
sulphate, and this in its turn with a portion of the mercury to form
mercurous sulphate. The hydrolysis of the other sulphites of
mercury to metal and acid, and of mercuric sulphate to oxide and
acid, gives every probability of the correctness of the assumption
here made respecting that of mercuric oxysulphite. As for the
secondary union of mercury with mercuric sulphate, that can be
readily effected in the cold, experimentally, in the presence of water
containing very little sulphuric acid. But the mercuric oxide will
also, when hot, act as an oxidising agent upon undecomposed oxy-
snlphite, and thus determine the explosion. In proof of this, we have
ascertained that a mildly explosive mixture can be made from pre-
cipitated mercuric oxide and either of the other mercury sulphites.

Water causes mercuric oxysulphite to decompose more rapidly than
it does when kept dry. Heated with water it decomposes, the change
heing at first probably the same as when it is heated dry; but the
water effects a further change, and the products are mercurous
sulphate, mercuric oxysulphate, sulphuric acid, and mercury. When
dissolved in aqueous solution of sulphurous acid and boiled, it decom-
poses in the manner which seems normal to a non-basic mercury
sulphite, the sole products being mercury and sulphuric acid.

Potassium hydroxide converts it into mercuric oxide insoluble, and
mercuric potassium sulphite in solution :—

Hg(SO,0Hg0),Hg + 2KOH = Hg(SO0,0K), + 3HgO + H.O.

548 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

Hydrochloric acid, added in excess at once, converts it into mercuric
chloride and sulphurous acid. Hydrochloric acid, graduaily added,
causes the change of part of it into mercurosic and even hypomercu-
rosic sulphite, by the sulphurons acid liberated from another part ;
more hydrochloric acid then forms mercurous chloride. If it has
changed at all into mercurous sulphate, some mercurous chloride and
sulphuric acid will also be produced, but sulphurous acid, even in the
nascent state, has no action on mercuric chloride in the cold. If,
again, it contains any of its mother-liquor of mercuric nitrate, reac-
tions will take place on adding the hydrochloric acid, which will also
produce mercurous chloride and sulphuric acid, the liberated sul-
phurous acid forming mercurosic sulphite from the nitrate, and the
mercurosic sulphite forming mercurous chloride with the mercuric
chloride. (See “ Mercurosic Sulphite,” p. 563.)

Nitric acid, or even sulphuric acid, if sufficiently dilute, hardly
dissolves it, but hastens its change to mercurous sulphate. When a
little more concentrated, the nitric or sulphuric acid dissolves it,
without liberating sulphur diowide, and then, if hydrochloric acid is
added to the solution without delay, mercuric chloride and sulphur
dioxide are formed in considerable quantities, besides some mercurous
chloride and sulphurous acid.* But if the solution is kept for a few
minutes, mercurous sulphate takes the place of the mercuric oxy-
sulphite, and soon begins to separate. The addition of hydrochloric
acid now produces only mercurous chloride and sulphuric acid.
Sufficiently strong sulphuric or nitric acid liberates sulphur dioxide
at once from mercuric oxysulphite.

Sulphurous acid, free or nearly free from sulphuric acid, when
added in excess at once, dissolves it without forming any sulphate or
sulphuric acid, and leaving mere traces of metallic mercury, the
result of previous change in the oxysulphite; when, however, the sul-
phurous acid is added gradually, mercurosic sulphite is formed, then
hypomercurosic sulphite, and, lastly, metallic mercury, as the quantity
of sulphurous acid increases, while the solution from the first contains
sulphuric acid, and, for a time, a little (acid) mercuric sulphite. The
precipitation of mercurosic sulphite is also quickly induced when the
sulphurous acid is added at once, provided it contains a good quantity
of sulphuric acid. The later changes, here described, are treated of
in subsequent sections of this paper; the formation of the mercurosic
sulphite is represented by the equation—

3(OHg.SO;). + 6SO,;H, = 4S0,H, + 4Hg;(SO;). + 20H;.
The effect of boiling the solution of mercuric oxysulphite in sul-
phurous acid has already been stated.

* Sulphuric acid can, of course, only be detected as formed, when it has not been
used.

AND THE CONSTITUTION OF SULPHITES. 549

Sodium sulphite solution, added quickly in excess, dissolves it as
mercuric sodium sulphite; a minute quantity of metallic mercury
is left, due to the presence of a little mercurous sulphate formed
by previous change of the oxysulphite. Sodium hydroxide is also
formed :—

(OHg.SO,), + 6Na(SO;Na) + 20H, = 40HNa + 4H¢(SO,Na)>.

When the sodium sulphite is added gradually, however, black, flocculent
hypomercurosic sulphite is first seen, which is subsequently resolved
into mercury and mercuric sodium sulphite. This production of hypo-
mercurosic sulphite seems to prove that when only partially attacked
by sodium sulphite, the mercuric oxysulphite is decomposed in such a
way that the basic oxide is removed, and the normal sulphite left to
suffer a change practically the same as hydrolysis, but effected by the
sodium hydroxide which has been formed along with it, instead of by
water. Thus :—

2H¢(SO.0).Hg,0, + 8Na(SO,Na) + 40H, = 2Hg(SO,0)Hg
+ 4Hg(SO,Na), + 8NaOH.

2He(S0.0),Hg + 4NaOH = Hg(SO,0),Hg, + 2Na,SO, + 20H».

Silver sodium sulphite solution dissolves mercuric oxysulphite, all
the silver being precipitated, principally as sulphite. The precipitate
is much discoloured, apparently by the presence of a little mercuric
argentous sulphite (coming from mercurous sulphate). Silver oxide
scarcely appears, because the original solution always contains some
free sodium sulphite, and accordingly sodium hydroxide is found in
solution instead,

Sodium chloride solution forms mercuric oxide, yellow and insoluble,
and mercuric sodium sulphite and mercuric sodium chloride, both in

solution :—
(OHg.SO;). + 4NaCl = 20Hg + Hg(SO;Na),. + HgCl,Na,.

A little mercurous chloride is found with the oxide, and a little
sodium sulphate in the solution, owing to the mercuric oxysulphite
having begun to change into mercurous sulphate by the time it is
prepared for experiment.

Potassium iodide solution changes mercuric oxysulphite instantly
into a dull red compound, soluble in excess of the reagent, but with
much more difficulty than simple mercuric iodide, and then yielding
an alkaline solution. This dull red substance is of complex nature,
and consists of mercuric iodide, oxide, and sulphite, in some state of
combination. Except when the solution of potassium iodide is very
concentrated, the dissolution of the red compound is at once followed

VOL. XUIx. 2p

550 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

by the appearance of a slight yellow precipitate convertible by a
stronger solution of potassium iodide into a minute quantity of
metallic mercury and dissolved mercuric iodide. The production
of this yellow iodide-is to be referred to the presence of some
mercurous sulphate in the mercuric oxysulphite.

Mercuric chloride. solution acts slowly on mercuric oxysulphite, if
at all, mercurous chloride being gradually produced, together with
mercuric sulphate which remains in solution with the excess of
mercuric chloride. -By dissolving mercuric oxysulphite in sul-
phuric acid before adding the mercuric chloride, the precipitation
of mercurous chloride begins immediately, and proceeds rapidly
until all the sulphite has been decomposed. Mercuric chloride has
probably no action on mercuric oxysulphite, its action being on the
mercurous sulphate into which the oxysulphite so readily changes.
The action of mereuric chloride .on other mercury sulphites is sharp
and unmistakable, and in their case the completion of the action is
retarded, instead of advanced by the presence of sulphuric acid.

Mercuric iodide, oxide, nitrate, and sulphate, and also mercurous
nitrate and sulphate are without action. Silver nitrate also is without
action on the oxysulphite itself, but a little silver is taken up
apparently through reaction with the mercurous sulphate present.

Formation and Preparation.—Mercuric oxysulphite is formed
in reactions between—

(a.) Solutions of sodium sulphite or mercuric sodium sulphite and
mercuric nitrate ; or solutions of sodium sulphite or mercuric
sodium sulphite, strongly alkaline, and of mercuric sulphate.

(b.) Silver sulphite and solution of mercuric nitrate.

a. It was prepared first from sodium sulphite and mercuric nitrate
by St.-Gilles, who states that other mercuric salts cannot be em-
ployed. In this statement he is practically correct, but by adding
a strong solution of potassium or sodium hydroxide to the solution of
sodium sulphite, this solution can be made to precipitate the oxy-
sulphite from mercuric sulphate and yet leave an acid mother-liquor.

Held in acid solution, the mercuric oxysulphite soon becomes
mercurous sulphate. If, without using alkali, the attempt is made to
neutralise sulphuric acid by adding more sodium sulphite, mercurosic
sulphite is precipitated. Mercuric oxysulphate cannot be used, as
from its insolubility sodium sulphite would, for a time, be in excess
in the solution.

Mercuric Nitrate and Sodium Sulphite.—Consult the account already
given of the reactions between these substances, and between mercuric
nitrate and mercuric sodium sulphite. The mercuric nitrate solution

ww FF

AND THE CONSTITUTION OF SULPHITES. 551

should not be too dilute; St.-Gilles used even a “syrupy” solution.
Such concentration is unnecessary, but there should be at least one of
nitrate in ten of solution. The sodium sulphite solution is better
when rather dilute; a 5 per cent. solution answers well. Both
solutions should be free from chloride, otherwise mercurous chloride
will be found with the oxysulphite. The sodium sulphite should be
in quantity small enough to leave much mercuric nitrate undecom-
posed, and be added to the nitrate all at once, and with stirring.
A larger yield of oxysulphite may be obtained by mixing sodium
hydroxide with the sodium sulphite, but the oxysulphite is then less
easily purified. The use of mercuric sodium sulphite in place of
sodium sulphite is an improvement, because although it causes the
liberation of as much nitric acid as the single sulphite, it generates
only half as much sodium nitrate for the same quantity of oxy-
sulphite.

Mercuric oxysulphite cannot be washed much without being
rendered impure by decomposition into mercurous sulphate, and the
mercuric nitrate cannot be all washed out of it by water alon,
because this precipitates basic nitrate; a little nitric acid must be
added to the water. Another difficulty in the way of washing the
precipitate is that it is far less stable in the moist state, when mer-
curic nitrate is absent. The process of purifying the oxysulphite
without washing by spreading it in a thin layer upon a good porous
tile, has proved very satisfactory, the mother-liquor being withdrawn
from it very thoroughly, and it is left as a dry, undecomposed
powder, when scraped from the tile after 20 or 30 minutes (but this
process is not satisfactory when, by using free alkali, mercury nitrate
has been too far replaced by sodium or potassium nitrate). When
well prepared, the oxysulphite should be of a pale-yellow colour,
dissolve in hydrochloric acid to an almost clear solution, and be very
explosive. In precipitating, it is for a moment curdy, and then
becomes granular.

b. The decomposition of mercuric uitrate by silver sulphite is a
very interesting fact, which depends largely, no doubt, on the relative
instability of oxylic mercury salts, such as the nitrate, in comparison
with the oxylic silver salt, but would seem from other evidence to be
partly due also to a firmer union of sulphuryl directly with mercury
than with silver: in the reaction, the silver takes the place of sodium.
The silver sulphite is added to excess of the mercuric nitrate with
trituration, when it is quickly replaced by mercuric oxysulphite as
the insoluble matter. As is the case when sodium sulphite is used,
only a part of the mercuric nitrate can be precipitated as oxysulphite ;
more than sufficient silver sulphite for this purpose, no doubt,
produces more oxysulphite, but this now remains dissulved in the

2P2

552 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

nitric acid which has been formed, and soon becomes mercurous
sulphate, which crystallises out. Prepared from silver sulphite,
mercuric oxysulphite is more distinctly yellow, and always retains a
very little silver,* which washing with mercuric nitrate solution does
not remove. No useful quantitative analysis therefore could be made.
It is highly explosive, slowly changes to mercurous sulphate, loses
water on exposure, and in all respects is identical with that otherwise

prepared.

Chemical Composition.—Mercuric oxysulphite is readily distin-
guished. Hydrochloric acid shows it to be a sulphite. Sodium hydr-
oxide and sodium cbloride both precipitate yellow mercuric oxide from
it, and thus show it to be a mercuric salt and a basic compound. Hydro-
chloric acid also shows it not to be a mercurous salt, and, therefore, not
one of the other sulphites of mercury. Its entire volatility, when pre-
viously moistened and then heated to nearly a red heat, shows the
absence of alkali salt. Its great explosibility when dry is not only quite
characteristic, but may be made a test of its freedom from alkali
salt, which even in very small quantity greatly lessens this quality.
Its gradual conversion into mercurous sulphate is evidence of its
ultimate composition, and the simultaneous decrease in weight, proof
of its containing water. The fume arising from its explosion also is
damp. Some conversion to mercurous sulphate cannot be avoided
during its preparation for analysis, but not necessarily to a greater
extent than to cause its hydrochloric acid solution to be somewhat
opalescent, and contain an insignificant amount of sulphate. Mercuric
oxysulphite differs much in appearance from the other mercury
sulphites ; it is of a pale-yellow colour and forms a dense precipitate,
at first curdy, then granular ; hypomercurosic sulphite on the contrary
is black and flocculent, whilst mercurosic sulphite forms a vividly
white, voluminous, sometimes crystalline, precipitate.

By attributing to the oxysulphite a constitution in which one-
fourth of its mercury is entirely in non-oxylic union with the
sulphuryl, its reactions are made most capable of receiving consistent
representation as well as its property of resisting the action of nitric
acid (see the section on the general constitution of mercury sulphites).
Although most mercuric oxysalts have three times the proportion of
mercury proper to the normal salt, the oxysulphite has only twice,
because it is only half an oxylic salt, and, as such, has only half as
much extra base as the wholly oxylic basic salts. This is shown by

* Apparently as silver sulphate. Prepared from sodium sulphite, the oxysulphite
takes up a very little silver when shaken with silver nitrate solution, and retains it
during washing. Mercurous sul) hate and silver nitrate do, we find, yield silver
sulphate and mercurous nitrate.

AND THE CONSTITUTION OF SULPHITES. 553

comparing the formule of the sulphite with that of the oxysulphate
(“‘ yellow turpeth mineral”) :—

Oxysulphite. Oxysulphate.
Sr O0Hg0. _ -0Hg0-S0,0Hg0
Hg<go,-oHgo> Hs: H8<0Hg0-80,-0Hg0> H8-

The results of St.-Gilles’s analyses of the oxysulphite were not
concordant, and led him to believe that it was mixed with more or
less normal sulphite. We analysed our preparations by dissolving
them in hydrochloric acid containing a little bromine, and then ex-
posing the solution to air, without heating, until it became colourless.
The sulphuric acid was precipitated first as barium salt ; and then the
mercury as sulphide, which was finally dried at 105—110°. Water was
estimated from the loss, first in an air-pump desiccator, and then in
an oven not heated above 60°. Exposure in the desiccator fcr
50 hours removed nearly all the water; the further exposure of
20 hours giving only an additional milligram loss. The salt now
contained much sulphate, but was still explosive. Exposure to a heat
of 55—60° for six hours destruyed the explosiveness of the salt, and
caused a further loss of a milligram (salt taken = 0°656 gram). This
loss may have been water, or it may have been mercury; we took it
as the former. Heated now for six hours at 120°, it lost 2 mgrms. ;
this loss we regarded as mercury, and testing the residue found a
little mercuric sulphate in it :—OHg.SO; = Hg + HgSOQ,. Analyses
I and II were made considerably earlier than III and IV, and in
very hot weather; the preparations were also exposed for a longer
time on the tile. Preparations III and 1V were sensibly dry, and
formed non-coherent powders. Loss of water by efflorescence and by
conversion to mercurous sulphate all tells on the mercury percentage
number, because mercury constitutes four-fifths of the salt. TII was
washed with water slightly acidified with nitric avid; the othe.s
were not washed. All four were different preparations. A portion
of the salt prepared for analysis IV was ignited to see to what
extent it might be impure from retained bye-products; the residue
amounted to U'17 per cent. :—

I I. UL IVa IVb. OsHg,(S0,),0H.
Mercury .. 79°64 79°81 7871 7860 — 79°21
Sulphur... 650 635 606 660 — 6°34
Oxygen... — — — — — 12°67
Water.... — —_ — — 186 1:78
100-00

554 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

Mercuric Hydrogen Sulphite unknown in the Separate State.

Haloid salts of mercury always give a mercuric alkali sulphite
when treated with an alkali sulphite ; but Wicke states that by treat-
ing solid mercuric chloride with a solution of sodium pyrosulphite he
prepared mercuric hydrogen sulphite. He is certainly wrong. A
solution of sodium pyrosulphite readily dissolves mercuric chloride in
powder, and, if used in excess, the solution soon deposits an abundant
crystalline precipitate soluble in much water. All so far is as Wicke
describes it ; but the precipitate, however, is mercuric sodium sulphite,
and no‘, as he supposed it to be, mercuric hydrogen sulphite. The
dissolution of the mercuric chloride liberates much sulphur dioxide,
which escapes with effervescence when the solution of pyrosulphite is
concentrated, as it should be, in order to precipitate the salt. There
should be, of course, no liberation of sulphur dioxide if Wicke were
right, nor should the crystalline salt yield a large residue of sodium
sulphate on ignition as it does. Wicke’s salt gave him a bright
yellow basic salt when treated with potassium hydroxide. The un-
dissolved crystals may, indeed, give a little red mercuric oxide with
solution of sodiwm hydroxide of great concentration, but not even
that with a solution of potassium hydroxide. Even if the crystals
were mercuric hydrogen sulphite, they would not give a basic pre-
cipitate with potassium hydroxide, but colourless, neutral, sparingly
soluble mercuric potassium sulphite. Only mercuric chloride or its
compound with sodium chloride would give a bright yellow precipi-
tate such as Wicke obtained.

What is probably a solution of mercuric hydrogen sulphite is
described in the following section.

Mercuric Hydrogen Sulphite Solution.

Three chemists have published observations on the reactions of
sulphur dioxide with mercuric oxide in the presence of water—
Rammelsberg, Péan de St.-Gilles, and Vogel; all fuund a white
substance take the place of the mercuric oxide, and another mer-
cury compound go into solution, but beyond this their observa-
tions do not agree. Rammelsberg found the white substance
to be a mercurous anhydrosulphite of varying composition, and the
compound in solution to be mercurous sulphate. St.-Gilles states that
the solution yielded a precipitate of a mixture of mercuric sulphite
and mercurous sulphate, and that until this precipitate had formed it
did not contain either a mercurous salt or a sulphate; whilst Vogel
found the white substance to be mercurous sulphate, partly soluble in
the mother-liquor as a compound not precipitable by common salt
( Watts’s Dictionary ; Gmelin’s Handbook).

AND THE CONSTITUTION OF SULPHITES.

We agree with none of these findings as a whole. We hold that
neither Rammelsberg’s mercurous anhydrosulphite nor St.-Gilles’
mercuric sulphite exists; we find, moreover, that Vogel’s white
“mercurons sulphate ” is really mercurosic sulphite. Coming now to
the solution obtained along with the white substance, it certainly
contains sulphuric acid from the very first appearance of the white sub-
stance, and St.-Gilles is wrong in denying Rammelsberg’s statement
to that effect. Further, it contains its mercury as mercuric salt, as
St.-Gilles and Vogel (?) seem to have held, and not as mercurous
salt, as Rammelsberg has stated.

None of these.chemists seem to have noticed that mercuric oxide
may be dissolved: in sulphurous acid without the formation of any
precipitate ; the conditions for seeuring such a dissolution of the
oxide are, however, simple. The sulphurous acid solution must not
contain much sulphuric acid (commonly present as the result of
atmospheric oxidation) ; and the mercuric oxide must be prepared by
precipitation, must be added well divided and suspended in water to
the sulphurous acid, and-only in small.quantity, because only a weak
solution can be prepared. A solution, the same in every respect, can
be obtained by using mercuric oxysulphite in place of mercuric oxide.
Mercurosic sulphite and hypomercurosic sulphite can also be used in
place of the oxide, but then there is a residue of metallic mercury. An
impure solution can be prepared by-cautiously adding, in small quantity,
a solution of either mercuric nitrate or sulphate, or of mercurous nitrate
or solid mercurous sulphate or.mercuric oxysulphate to sulphurous
acid; with mercurous salts, half the mercury separates insoluble, and
with all these salts either sulphuric or nitric acid remains in the solu-
tion. Even by the action of a little nitric acid on mercuric sodium
sulphite a dilute mercuric hydrogen sulphite solution can be
obtained.*

restr g

a
Le

r 30 £3

eS
Bra
t

“es

roy Sem ¢
be OW bu &
hs

Properties.—The solution reacts as one of mercuric sulphite in
sulphurous acid. It is more probable that it contains mercuric acid
sulphite than the normal sulphite, and may, therefore, be regarded as a
solution of that salt, in absence of any evidence to the contrary. The
acid sulphite is the hydrogen double salt, Hg(SO;H),, similar to the
sodium double salt, Hg(SO;Na)., in which the mercury is wholly
non-oxylic. Thus constituted, its ready conversion to mercurosic
sulphite admits of more satisfactory explanation than if its mercury
were half oxylic. The question as to the non-existence of acid
sulphites of the alkali metals is not affected by the existence of this
specially constituted mercuric salt.

* See also the account, in the latter part of this section, of the action of sul-
phurous acid on mercuric chloride.

556 DIVERS AND SHIMIDZU : MERCURY SULPHITES,

Unlike the mother-liquor of mercurosic sulphite, the solution of
mercuric hydrogen sulphite, when prepared from pure materials, con-
tains no sulphuric acid, moreover, all its mercury is in mercuric com-
bination. Alkalis give no precipitate of mercuric oxide, because they
form a soluble double sulphite, precipitable, however, by alcohol. Any
specific action on mercurous salts cannot be investigated, because of
that exerted on them by the sulphurous acid present in the solution.

Mercuric hydrogen sulphite solution is very unstable, and readily
suffers hydrolysis, either wholly or in part—wholly into mercury and
sulphuric acid, in part into mercurosic sulphite and sulphuric acid—
sulphurous acid remaining free in both cases. The solution deposits
all its mercury slowly at common temperatures, but at once when
heated, and an equivalent quantity of sulphuric acid remains in
solution.* The reaction may be expressed by the equation—

H,0 + Hg(SO;H), = Hg a HO-SO,;H + H:-SO,H.

Partial hydrolysis, by which mercurosic sulphite is formed, is
brought about in two ways. Sulphuric or nitric acid, added in small
but sufficient quantity, causes it. Secondly, it is caused by adding
more mercuric oxide or some mercuroxylic salt (best, a solution of
mercuric nitrate or sulphate) to the solution. Whether the addition
proves sufficient or not to exhaust all free sulphurous acid is a cireum-
stance which does not affect the precipitation. Success with such
different reagents renders it improbable that the hydrolysis depends
on any direct reaction between them and the mercuric hydrogen
sulphite. The explanation we offer of the change is that it is induced
by effecting a state of supersaturation of the solution with mercuric
hydrogen sulphite, so that partial hydrolysis occurs and mercurosic
sulphite is precipitated. Mercury sulphites are insoluble in dilute
nitric or sulphuric acid, and therefore the addition of one of these
acids brings about the state of supersaturation which leads to the
sudden hydrolysis and precipitation. Mercuric oxide or salts generate
more acid sulphite in the solution, and in this way cause supersatura-
tion. According to this explanation, mercuric hydrogen sulphite must
be regarded as a salt soluble in a large quantity of water, and hydro-
lysed by less water.

The result of the hydrolysis of mercuric hydrogen sulphite when
it has only proceeded to the extent of separating one-third of the
sulphur as sulphuric acid, appears to depend on the difference between
oxylic and non-oxylic mercury compounds; the mercuric radicle,
Hg, alone can hold the non-oxylic position in a sulphite, the

* Facts also mentioned by Vogel; but the solution he examined had been pro-
duced along with the white substance (mercurosic sulphite), and already therefore
contained sulphuric acid.

AND THE CONSTITUTION OF SULPHITES. 557

mercurous radicle, Hg,, cannot. The mercurous radicle, on the other
hand, can hold the oxylic position, provided a mercuric radicle is in
non-oxylic relation to the sulphuryl, whilst the mercuric radicle
cannot do so, except in the partial and complex way seen in the basic
mercuric oxysulphite. Accordingly, the conversion of mercuric
hydrogen sulphite to mercurosic sulphite may be represented by the
following equation, in which the upper line shows the hydrolysis of a
third of the sulphite, and the lower the double decomposition whereby,
with the assistance of mercury set free by hydrolysis, two molecules
of mercury hydrogen sulphite become mercury sulphite and hydrogen
sulphite :—
Hg’"(SO;H), + OH, }={ H:SO;H + HO-SO;,H

Hg’ (SO;H), + Hg"(SO,H). 2H-SO;H + Hg"(SO,),He’s.

Mercur‘c Chloride and Sulphurous Acid.—It is well known that
a nearly boiling solution of mercuric chloride is converted by a stream
of sulphur dioxide into mercurous chloride, hydrochloric acid, and
sulphuric acid. This reaction does not take place at all in the cold,
but the two substances are, even then, not indifferent to each other.

Mercuric chloride is very much more soluble in sulphurous acid
than in water. The solution on exposure to air rapidly crusts over
and also deposits much crystalline precipitate, both consisting of mer-
curic chloride rendered insoluble by the escape of some of the sulphur
dioxide. Mercuric chloride loses its greater solubility in sulphurous
acid if much sulphuric acid is present, so that the addition of dilute
sulphuric acid to the sulphurous solution quickly induces crystallisa-
tion of mercuric chloride. When saturated with mercuric chloride,
sulphurous acid dissolves much more mercuric owide than it can other-
wise do, without yielding an immediate precipitate. Kither on
standing or on addition of sufficient mercuric oxide, precipitation
begins and slowly proceeds, the precipitate being crystalline mercurous
chloride. The appearance of this is equivalent to that of mercurosic
sulphite, for, as is described in the next section, this sulphite and
mercuric chloride at once react yielding mercurous chloride and
mercuric sulphite. Sulphuric acid is formed along with the mercurosic
sulphite. That the solution slowly decomposes, even when not fully
saturated with mercuric oxide, may be accounted for by regarding
a solution of mercuric chloride in sulphurous acid as being charged
with mercuric hydrogen sulphite, formed by decomposition of some
of the mercuric chloride :—

HgCl, + 2H(SO,H) = Hg(SO,H), + 2HCl.

As such it will, by dissolving mercuric oxide, become supersaturated
with mercuric hydrogen sulphite, through the action of the oxide on

558 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

the free sulphurous acid, even though still more oxide may be soluble
in it, on avcount of the hydrochloric acid present. Mercuric chloride
is also much more soluble in a solution of sodium sulphite or pyro-
sulphite, and is then certainly converted into mercuric sodium
sulphite.

In further support of the belief, expressed in the last paragraph,
that mercuric chloride and sulphurous acid partly change into mer-
curic hydrogen sulphite and hydrochloric acid, it may be pointed out
that, according to Sartorious, it is only between 70° and 80°, and in
dilute solutions, that mercuric chloride is completely precipitated as
mercurous chloride by sulphurous acid: for, in the first place, it is
just at that temperature that mercury sulphites suffer hydrolysis,
mercuric sulphite then passing into the stage of mercurosic sulphite,
from which mercurous chloride would be precipitated by the hydro-
cbloric acid; and, in the second place, too concentrated a solution of
mercuric ch!oride would, in the course of precipitation, charge the
water with so much sulphuric acid that it would stop the conversion
of the remainder of the mercuric chloride into sulphite, and therefore
into mercurous chloride. The action of sodium sulphite on mercuric
chloride also supports the above view, for here the precipitation of
the mercurous chloride by heat is certainly preceded by the formation
of mercuric sodium sul!phite (see the sections on “ Normal Mercuric
Sulphite and Mercuric Sodium Sulphite’’).

It may be startling to hear of sulphurous acid decomposing mer-
curic chluride, when sulphuric acid cannot do so, but then it must be
remembered that sulphurous acid completely decomposes both the
mercury sulphates, as well as the nitrates, into sulphites and free
acid.

Mercurous Chloride and Sulphurous Acid.—According to Vogel,
mercurous chloride is converted by sulphurous acid into a grey sub-
stance, probably a subchloride; we find the action to be exceedingly
slight, however. Ordinary calomel becomes a little greyish when
suspended in a small quantity of water through which sulphur
dioxide is passed; the effect is soon produced, and does not increase
during continued contact with the gas: only traces of hydrochloric
acid are to be found in the mother-liquor; there are also traces of
sulphuric acid, but these would soon be formed in any case in the
solution of sulphurous acid. Left for a day in old solution of sul-
phurous acid, calomel became dark grey, although even then the
action was but slight, for very little hydrochloric acid was in solu-
tion. In this case, the sulphuric acid which had accumulated in
the solution, probably assisted by rendering the calomel slightly
soluble. Precipitated mercurous chloride becomes coloured more
readily in water containing sulphurous acid, but still only light grey

AND THE CONSTITUTION OF SULPHITES. 559

and with the formation of mere traces of .hydrochloric acid. The
difficulty of getting precipitated mercurous chloride quite free from
other salts makes this slight reaction of little significance. We
doubt whether there is any reaction at all between true mercurous
chloride and sulphurous acid. When mercurous chloride is preci-
pitated by sulphurous acid itself, it is of dazzling whiteness, which is
a strange fact, if it is discoloured by sulphurous acid in other cases.
(For an account of the behaviour of mercurous chloride with sodium
sulphite, see the section on “‘ Hypomercurosic Sulphite,” p. 572.)

Mercurosic Sulphite.

Mercurosic sulphite was in the hands of both Rammelsberg and
St.-Gilles, without its true nature being recognised by either of them.
It is here, for the first time, therefore, described at any length, and
with the composition we give to it. It is a well-defined salt, although
from its history it might be supposed not to be such.

Properties.—Mercnrosic sulphite is obtained in somewhat dif-
ferent states. Prepared by adding mercuric oxide to sulphurous acid
solution, it is in lustrous, crystalline particles, which form felted sheets
like paper, when drained on a tile till dry. Prepared from a paste of
mercuric oxide and gaseous sulphur dioxide, it is not visibly crys-
talline, but it forms a voluminous precipitate of crystalline habit, is
brilliantly white, and, when pressed whilst moist, crepitates like
starch. Prepared by reaction between another mercury sulphite and
a mercury nitrate or sulphate, it is a voluminous, apparently amor-
phous precipitate, of a buff colour. It is then not quite pure, indeed,
but its colour can hardly be referred to the presence of any coloured
impurity. Since mercurous chloride occurs both of a fine white
colour and of a pale buff, it seems allowable to hold that mercurosic
sulphite may alsodo so. We have not got the buff-coloured variety free
enough from sulphate, for useful quantitative analysis, but its quali-
tative reactions are those of the white mercurosic sulphite.

Mercurosic sulphite is insolubie in water. Its composition is ex-
pressed by the formula Hg;(SQ;),(OH,),. It is very efflorescent, and
becomes anhydrous in the desiccator. Only when it is crystalline
can it be obtained dry, with all the water of composition retained.
Rammelsberg found in his red, crystalline cuprosic (or ‘ cuproso-
cupric”) sulphite 2 mols. H,O, and St.-Gilles in his yellow amor-
phous cuprosic sulphite, 5 mols. H,O. Our preparations of mercurosic
sulphite, air-dried in hot weather, have shown slightly less than
2 mols. H,O, and we cannot be certain that our preparation having
4 mols. H,O had not already effloresced, and lost a fifth mol. H,0,
although we believe it had not.

560 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

Mercurosic sulphite is so stable when dry as to be but very little
changed after long keeping. In a closed vessel, it gradually darkens
in colour, but nearly bleaches again on exposure to the air for some
time. ‘he darkening is in all probability due to the formation of a
little hypomercurosic sulphite. Slowly in the air, more quickly in a
vacuum, it acquires a slight permanent grey colour, due to the pre-
sence of a minute quantity of mercury. The slow decomposition of
mercurosic sulphate results in the production of mercurous sulphate,
metallic mercury, sulphur dioxide, and water, the same products as
those produced by heat. Sulphurous acid blackens both mercurous
sulphate and mercurosic sulphite by forming hypomercurosic sul-
phite ; hence, no doubt, the partial blackening when the mercurosic
sulphite is keptshut up. After blackening and bleaching in the air, the
changed sulphite may be somewhat buff-coloured.

Heated dry at a temperature of about 80°, it evolves sulphur
dioxide, mercury, and water, and becomes at first black, then grey-
white, leaving mercurous sulphate and some of the mercury as a
residue; these may be separated by further heating, so as to vola-
tilise the mercury. The blackening may certainly be referred to
the formation of hypomercurosic sulphite, and although it is for the
time very great, it is only transitory if the heat is maintained.
Examined when blackest, the amount of undecomposed sulphite
proves to be exceedingly small. It is not probable that hypomer-
curosic sulphite forms an intermediate product in the decomposition
of the mercurosic sulphite; the moist sulphur dioxide no doubt pro-
duces it by a secondary action on the undecomposed mercurosic
sulphite, or on the mercurous sulphate (see ‘“ Hypomercurosic
Sulphite,” p.571). The decomposition of mercurosic sulphite by
heat is represented by the equations—

Hg;(SOs)2 + 20H, => dHg - 2S0,H2.
Hg + Hg,(SO,), + 280,H, = 2Hg,SO, + 280, + 20H).

Some quantitative measurements of this reaction have sometimes
proved in close agreement with the equation, and never widely dif-
ferent from it. We have got 61°2, 584, 56:1, and 61°9 per cent. of
mercurous sulphate. With 2 mols. H,O retained by the salt used, the
calculated number is 62°3 per cent.; with 4 mols. H,O nearly 60 per
cent. The sulphur dioxide was also measured, but as the gas was not
dried, the results were only rough approximations. Instead of 8 per
cent., 8°6, 7°8, 8°5, and 8°l were found. The experiments were made
with the aid of the Sprengel pump. There are two material sources
of loss of mercurous sulphate in such an experiment ; the mercurous
sulphate is apt to be lost as fume; and the hydrolysis of the sulphite
is liable to proceed more rapidly than the action on the remaining

AND THRE CONSTITUTION OF SULPHITES. 561

sulphite of the sulphuric acid generated by the hydrolysis. A low
production of sulphate should be more evident when the water is as
much as 4 mols., and this probably accounts for our getting such a
low number as 56 per cent. in one case.

Heated with water, it is converted tumultuously into metal and
sulphuric acid.* Enough water being present, no mercurous sul-
phate is produced :—

Hg;(SOs). + 20H, a 3Hg + 2S0,H:2.

When the salt is heated dry, its decomposition begins in all pro-
bability in this way, but is then modified by the reaction between the
sulphuric acid and unchanged sulphite, giving mercurous sulphate
and sulphurous acid. The non-production of any mercuric sulphate
is to be expected, as it is easy to show experimentally the instant
conversion of mercuric sulphate to mercurous sulphate by moist mer-
cury.

Potassium hydrowide converts it into mercurous oxide, insoluble,
and mercuric potassium sulphite, going into solution—

the potassium displacing the mercurous but leaving the mercuric
radicle.t

Hydrochloric acid at once decomposes it, the products being sul-
phurous acid and the two chlorides of mercury. When the mercurosic
sulphite is buff-coloured, the mercurous chloride from it is of the
same colour. No sulphuric acid is produced in this reaction. The
preparations of buff-coloured sulphite, however, to start with,
always contain a small quantity of sulphate. Nitric acid and sul-
phurie acid in the dilute state are without action on mercurosic sul-
phite. Stronger acids dissolve it, but the nitric acid scarcely before
being strong enough to oxidise the sulphite. Sulphurous acid blackens

* The property was observed by Rammelsberg in what he regarded as mercurous
anhydrosulphite, which he prepared in the same way as is here given for preparing
mercurosic sulphite.

+ Rammelsberg, by similar treatment of the salt prepared by him, obtained very
little potassium sulphite. He would seem to have been working then with a pre-
paration mostly mercurous sulphate. According to Watts (Watts’s Dictionary) he
found his salt to “ oxidise” rapidly to mercurous sulphate, mercury becoming free.
It is material to state that mercury sulphites do not oxidise at all: they do become
sulphates, but not by uniting with atmospheric oxygen. It is further to be remarked
that the salt, viewed as either mercurous anhydrosulphite or mercurosic sulphite,
could not well yield metallic mercury in becoming mercurous sulphate by oxidation,
because it already contains less than two atoms of mercury to one of sulphur. It is
rightly stated to yield mercury by changing into mercurous sulphate, but it does so
in the way mentioned in the text.

‘

562 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

and decomposes it, dissolving out the elements of mercuric sulphite
and leaving hypomercurosic sulphite insoluble for a time, but then
also decomposing (see section on ‘‘ Hypomercurosic Sulphite,” p. 567).
The presence of sulphuric or nitric acid greatly hinders, and for a
time at least altogether prevents the action of the sulphurous acid, a
circumstance greatly facilitating the preparation of mercurosic sul-
phite. The influence of these acids is important, not only in preserving
mercurosic sulphite, but also in connection with its production (see
the section on “ Mereuric Hydrogen Sulphite Solution,” p. 554).

The reaction between sulphurous acid and mercurosic sulphite in
absence of sulphuric acid, consists in the exchange of non-oxylic
hydrogen for oxylic mercury, as expressed in the following equa-
tion :—

2H¢(SO;),Hg, + 2H(SO,H) = Hg(SO,),Hg, + 2Hg(SO,H)..

Sodium sulphite solution produces effects similar to those of sul-
phurous acid, but acts more rapidly than the acid. The mercurosic
sulphite is blackened by being converted into hypomercurosic sul-
phite insoluble, and mercuric sodium sulphite dissolving ; here also
as with sulphurous acid, the sodium sulphite exchanges its non-oxylic
sodium for half the mercury of the oxylic mercurous radicle of the
mercurosic sulphite, the other half of this mercury going to form
hypomercurosic sulpbite with mercurosic sulphite :—

2Hg(SO;).Hg, + 2Na(‘$O,Na) = Hg(SO,;),.Hg, + 2Hg(SO,Na),.

The hypomercurosic sulphite thus formed is acted on by more
sodium sulphite, becoming mercuric sodium sulphite and free mer-
cury (see the section on “ Hypomercurosic Sulphite,” p. 569). Silver
sodium sulphite solution yields mercuric sodium sulphite in solution,
all the silver being -precipitated if not in excess. The insoluble
sulphite is very black, and appears to be a mercury silver sulphite ;
it is too black for hypomercurosic sulphite, even unmixed with white
silver sulphite ;* besides, hypomercurosic sulphite at once reacts with
silver sodium sulphite. The reaction between the silver sodium
sulphite and mercurosic sulphite appears, then, to be expressed by
the equation—

Hg(SO,),Hg, + 4Ag(SO,Na) = 2Hg(SO,),Ag, + 2Hg(SO,Na):,

in which mercuric argentous sulphite appears.
Sodium chloride solution added freely produces much black hypo-
mercurosic sulphite, which slowly disappears on digestion. By add-
* The “ silver sulphide” observed by Vogel, and again by Pleischl, to be formed

by the action of sulphurous acid and sulphites upon silver salts, must have been
argentous sulphite, and not sulphide at all.

~~ wa te ~

AND THE CONSTITUTION OF SULPHITES. 563

ing the sodium chloride gradually, hypomercurosic sulphite is only
produced by the first portions. This difference is due to the mercury
salts which go into solution by the action of the first portions modify-
ing that of those added afterwards. The final result of the addition
of sodium chloride is expressed by the equation—

Hg(SO;).Hg, + 2NaCl = Hg(SO;Na), + (HgCl).,

but the primary action appears to be that of an exchange of sodium
for the oxylic mercury, in which only half of this mercury combines
with the chlorine, the other half then going to unite with more mer-
curosic sulphite—

HgCl,Naz.

Then, as the quantity of mercuric chloride increases, and the mercu-
rosic sulphite becomes less, all the mercury goes to form mercurous
chloride, while any hypomercurosic sulphite previously formed is
slowly converted into chloride (see the section on “ Hypomercurosic
Sulphite”). Mercurie sodium chloride acts in the same way as sodium
chloride, the mercuric chloride taking no part; it has no action of its
own on mercurosic sulphite, but when sodium chloride is present this
salt acts first. Mercuric chloride solution added, not in excess, pro-
duces mercurous chloride and sulphuric acid, all the mercury being
precipitated—

but if added at once in excess, much mercuric sulphite goes into solu-
tion in the mercuric chloride, and then slowly changes to mercurous
chloride and sulphuric acid, as already described under normal
mercuric sulphite. Potassium iodide solution forms a brownish-yellow
mercury iodide and a solution of mercuric potassium sulphite,
Mercuric iodide, oxide, nitrate, and sulphate, aud mercurous nitrate and
sulphate, are without action.

Formation and Preparation.—Mercurosic sulphite is formed in
reactions between—

(a.) Sulphurous acid, water, and mercuric oxide ;

(b.) Sulphurous acid, water, and mercuric oxysulphite ;

(c.) Sulphurous acid, water, and either mercuric oxysulphate or solu-
tion of mercuric nitrate, or solution of mercuric sulphate ;

(d.) Mercurie sudium sulphite and either dilute nitric acid or dilute
sulphuric acid (solutions of either sodium sulphite or mercuric
sodium sulphite, and either mercuric nitrate or mercuric sul-
phate) ;

564 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

(e.) A warm solution of mercuric sodium sulphite and mercuric
oxide ;

(f.) Mercurie sodium sulphite and either mercurous sulphate or a
solution of mercurous nitrate ;

(g.) Hypomercurosic sulphite and a solution of either mercuric nitrate
or mercuric sulphate.

a. (I.) Precipitated mercuric oxide made into a thick paste with
water is treated with a stream of sulphur dioxide in a vessel immersed
in cold water; the paste rapidly thins by dissolution of the oxide,
mercurosic sulphite being formed, and gradually increasing in quan-
tity. In this action, the oxide goes through the stage of soluble
sulphite in changing into mercurosic sulphite. Excess of sulphur
dioxide does no harm. If the mixture is allowed to grow hot, or if
too much water has been taken, there will be some blackening of the
product during the passage of the gas, due to the formation of hypo-
mercurosic sulphite; there will be blackening also if water is after-
wards added to the sulphite and its mother-liquor. The reason of this
is that the permanence of the mercurosic sulphite in presence of sul-
phurous acid is conditional on there being enough sulphuric acid in
the solution; this is a bye-product of the reaction, and the addition
of water acts injuriously by diluting this acid too much. Just at
first, in the preparation, before any sulphuric acid is formed, the
sulphurous acid is at once absorbed by the abundant mercuric oxide;
at this period of the operation, however, blackening occurs readily,
and the vessel should be well agitated until enough sulphuric acid
has been generated. Although the mercurosic sulphite can generally
be washed with strongly acidulated water without turning grey, it is
better to drain the sulphite on a good porous tile; in half an hour or
so it will be found to be dry, surprisingly free from sulphuric acid,
pulverulent and non-coherent when scraped off the tile, and almost
pure.

(II.) Precipitated mercuric oxide in paste is treated with sul-
phurous acid solution containing a moderate quantity of sulphuric
acid. This method of preparation yields the crystalline variety of
the sulphite. When drained dry on the tile, it forms glistening paper-
like laming. It is possible to prepare mercurosic sulphite without
blackening it, by the use of sulphurous acid unmixed with sulphuric
acid; to ensure success, the mercuric oxide should be in thick paste,
and the sulphurous acid concentrated and added gradually with
stirring and coo'ing.

b. Mercuric oxysulphite may be used instead of mercuric oxide in
the preceding method of preparation.

c. Mercuric oxysulphate may also be used in place of mercuric

ric

AND THE CONSTITUTION OF SULPHITES. 565

oxide in method a. Solution of mercuric nitrate, with the least
sufficient excess of nitric acid, is a very convenient source of (crys-
talline) mercurosic sulphite, when treated with sulphur dioxide,
either gaseous or in solution. A solution of mercuric sulphate, not
too acid, may also be taken.

d. Mercuric sodium sulphite and nitric or sulphuric acid. The
sulphite is best used in the solid state, and the suitable quantity of
acid added at once. Insufficient acid causes blackening. The reaction
is discussed in the section on “ Mercuric Sodium Sulphite ” (p. 540).
Mercuric sodium sulphite and mercuric nitrate form mercurosic
sulphite only by secondary reactions due to the nitric acid (see the
section on “ Mercuric Sodium Sulphite ”). Sodium sulphite and mer-
curic nitrate first form mercuric sodium sulphite and sodium nitrate, and
then the mercuric sodium sulphite reacts with more mercuric nitrate.
Were it not for the free acid of the mercuric nitrate solution, it
should be possible to add the nitrate until all sodium sulphite had
become mercuric sodium sulphite, and after that, further addition of
nitrate would form mercuric oxysulphite. But the quantity of free
acid sufficient to start the formation of mercurosic sulphite is very
small, and (neglecting to account for this) the following equation
serves to express the result of the change in presence of acid :—

3Hg(NO,), + 3Na,SO, + OH, = Hgi(SO,), + Na,SO, +
4NaNO, + 2HNO,.

e. Mercuric sodium sulphite and mercuric oxide react together as
described in the section on this sulphite.

f. Mercuric sodium sulphite and mercurous nitrate yield the buff-
coloured variety of mercurosic sulphite.

g. Hypomercurosic sulphite digested in a solution of mercuric
nitrate (or sulphate) quickly changes to the buff-coloured variety of
mercurosic sulphite. Hypomercurosic sulphite, as prepared, is nearly
always mixed with metallic mercury, but this does not remain in the
mercurosic sulphite prepared from it, as mercury is remarkably
soluble in mercuric nitrate solution.

Chemical Composition.—The action of hydrochloric acid clearly
shows mercurosic sulphite to be a sulphite, and both a mercuric and mer-
curous salt. The mercurous chloride produced could not come from the
mercuric chloride because sulphurous acid has no reducing action upon
itin the cold. It could indeed be formed by using an insufficient quan-
tity of hydrochloric acid at first, when the sulphurovs acid would act on
the remaining mercurosic sulphite (supposed here wholly mercuric)
and generate hypomercurosic sulphite, which with more hydrochloric
acid would give mercurous chloride; but the suggestion of any

VOL. XLIX. 24

566 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

such action is precluded by the fact that no sulphuric acid is formed
by the hydrochloric acid. Further, potassium hydroxide gives mer-
curous oxide; and sodium chloride gives mercurous chloride. The
mercuric chloride formed in abundance by hydrochloric acid, with-
out liberation of mercury, is full proof of its mercuric nature. Not
only is it easily distinguished from mercurous sulphate, by yielding
mercuric chloride and sulphur dioxide, and not sulphuric acid, but
also at once by decomposing in boiling water, not to yield a yellow
residue nearly as large as itself, but only grey mercury of exceedingly
small volume. Moreover, its appearance is characteristic. The presence
of water in it is readily detected by heating it in a tube. The repre-
sentation of mercurosic sulphite as having one-third of its mercury in
non-oxylic union with the sulphury]l, or as having the mercuric and not
the mercurous radicle so related is in full agreement with its reactions.

Sulphur and mereury were determined in the usual way, as already
d:scribed in the section on “Mercuric Oxysulphite.” Water was
estimated by drying in a vacuum over sulphuric acid; a few hours
proved sufficient to remove nearly all the water, but the time in the
desiccator was from 20 to 40 hours. A crystalline preparation
showed 4 mols. H,O, but other preparations showed less. These had
to be longer on the tile to get them dry, and loss of water by efflore-
scence took place; efflorescence proceeds so rapidly as to make the
salt show loss of weight during the operation of weighing.

I. 0°8005 gram taken for the mercury, and 0°8110 gram for the
sulphur determination.

II. 14955 grams taken for water.

III. 1:2533 grams taken for water.

IV. 1:2808 grams taken for water, and of the dried salt, 1:1413
grams for sulphur and nercury. All four quantities were taken from
separate preparations.

14H,0. I. Il. II. IV. He;(S0,)3,4H,0.
Mereury .. 7624 7594 — — 7220 72°12
Sulphur... 813 827 — — 804 7°69
Oxygen... 1220 — — — — 11°54
Water.... 343 — 321 503 859 8°65

—_— —_

100-00 10000

Rammelsberg’s Mercurous Anhydrosulphite believed not to Evxist.

Rammelsberg states that the mercury sulphite obtained by him, by
treating mercuric oxide with sulphur dioxide in the presence of
water, had the composition of a mercurous anhydrosulphite—either
(Hg.SO;),.SO, or (Hg.SO;),SO,. His results are neither very definite

AND THE CONSTITUTION OF SULPHITES. 567

nor theoretically very probable, and therefore as we have failed to get
in this way any other sulphite than mercurosic sulphite in the solid
state—a salt neither anhydrous nor exclusively mercurous—we are
convinced that he must, from some cause, have mistaken the nature
and composition of the product he obtained.

Hypomercurosic Sulphite.

Oxygenous mercurons salts are blackened by sulphurous acid and
soluble sulphites; they are then converted into hypomercurosic
sulphite, a well-defined salt, which has hitherto escaped recognition.
It was mistaken by Vogel for mercury, notwithstanding its black
colour.

Properties.—Hypomercurosic sulphite is a greyish-black amor-
phous substance, obtained as a voluminons, flocculent precipitate, inso-
luble in water. It is very unstable in water, but can be preserved dry
for a considerable time, without much change. Its composition is
expressed by the formula Hg,(SO;),0H,. Left to itself at common
temperatures, it very slowly evolves sulphur dioxide, leaving mer-
curous sulphate and mercury. The co-operation of moisture is, no
doubt, essential to this change, the primary form of which will be
hydrolysis into mercury and sulphuric acid, to be followed by reaction
between unchanged sulphite and the sulphuric acid. The production
of sulphur dioxide and mercury shows the change not to be one of
oxidation of sulphite to sulphate by the air. The equation of the
change is—

Hg,(SO;),0H, = 2Hg + Hg,SO, + SO, + OF;.

Heated dry, to a temperature of about 80°, it rapidly undergoes the
same change as that which takes place slowly in the air. On heating
it gradually in a vacuum, so as to volatilise the liberated mercury,
weights of mercurous sulphate were obtained equal to 46°90 per cent.
in one case, and 49°86 per cent. in another, calculation requiring
50°72 per cent. for Hg,(SO;),OH,. These results are satisfactory,
when it is considered that sulphuric acid may be carried off with the
mercury and water vapour, without acting on the remaining sulphite,
the decomposition being a rapid one; and, further, that in conse-
quence of previous hydrolysis during the operation of getting the
sulphite dry for use, a little mercury is liable to be already present
with the prepared sulphite.

Heated with sufficient water, it is rapidly and tumultuously con-
verted into metal and sulphuric acid :—

Hg.(SO;),0H, + OH, = 4Hg + 280, Hp.
2Q2

568 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

If the proportion of water is so small as to leave the sulphuric acid
somewhat concentrated, some mercurous sulphate is also formed, but
not otherwise. The production of the sulphate is then due to
reaction between unchanged sulphite and the sulphuric acid. We
must anticipate here, in order to say in explanation that hypomer-
curosic sulphite resists the action of rather dilute sulphuric acid. We
made an estimation of the mercury and sulphuric acid produced by
the hydrolysis of the sulphite, by weighing the former directly and
the latter as barium salt. The results, although not obtained under
the best conditions, may be given; they were :—Mercury, 82°33 per
cent., instead of 81°80; sulphur, 5°23 per cent., instead of 6°54.
Potassium hydroxide solution converts it into mercuric potassium
sulphite, which dissolves, and mercurous oxide and mercury which
remain insoluble. The decomposition of the sulphite by alkali is
expressed by the equation—
Hg(SO,0).Hg;,0H, + 2KOH = Hg(SO.0K).,0H, + Hg +
Hg.O + OH).

Hydrochloric acid at once liberates sulphur dioxide, and forms
mercuric chloride in some quantity. The insoluble matter only
slowly changes from the nearly black colour of the sulphite to a light
grey, even with thorough trituration of the whole in a mortar. It
then consists cf mercurous chloride and a very little free mercury,
while the soluticn contains no more than minute quantities of
mercuric chloride. The primary reaction is evidently expressed by

the equation—

The mercury and the mercuric chloride then slowly unite and form
more mercurous chloride, as they can be shown experimentally to have
the power to do. Possibly a black hypomercurous chloride, Hg;Cl,, is
one of the primary products, and then acts upon the mercuric
chloride :—

Hg;Cl, + HgCl, = 2(HgCl)>.

Nitric acid and sulphuric acid, when dilute, have no action. It
only reacts with nitric acid when that acid is strong enough to
oxidise it. Its resistance to acids is very striking, and furnishes a
ready means of quickly distinguishing it from mercurous oxide.
Acid solutions of mercurous salts are instantly precipitated by sulphur
dioxide, gaseous, or in solution. Sulphurous acid solution decomposes
it, producing a solution of mercuric hydrogen sulphite, and a residue

of mercury :—

Hg(SO,),Hg, + 2H(SO;H) = 2Hg + 2Hg(SO,H)..

AND THE CONSTITUTION OF SULPHITES. 569

There is here double decomposition, in which mercury in the oxyl'c
position exchanges with hydrogen into the non-oxylic one. Two-
thirds of this oxylic mercury become free, however, because only
mercuric hydrogen sulphite can exist. The action of sulphurous acid
on hypomercurosic sulphite is almost entirely prevented by the pre-
sence of some sulphuric acid; but for this, the sulphite could not be
prepared in a state at all approaching purity. The same is true of
mercurosic sulphite.

Sodium sulphite solution converts it into mercuric sodium sulphite
dissolving, and a residue of mercury, the action being similar to that
of sulphurous acid. Silver sodiwm sulphite solution takes mercury in
place of its silver, all the silver being precipitated, if not in excess.
The blackish hypomercurosic sulphite is converted into something
still blacker, probably mercuric argentous sulphite. Excess of silver
sodium sulphite does not destroy this black matter, or only slowly
destroys it. If boiled with water, it yields spongy silver amalgam ;
and when washed with cold water, it slowly undergoes the same
change. The reaction is perhaps—

Hg(S0;),Hg, + GAg(SO,Na) = Hg(SO,),Ag, + 2Ag +
3Hg(SO;Na)..

If the silver sodium sulphite is used in excess, and after the re-
action, it is diluted with water, it shows a greyish-white turbidity,
through the liberation of a minute quantity of silver. Apparently,
a very little mercuric argentous sulphite dissolves in the undiluted
solution of the other sulphites, and on dilution parts with some of its
silver.

Sodium chloride solution produces mercurous chloride, mercuric
sodium sulphite, and mercury :—

Hg(SO;),Hg, + 2NaCl = Hg(SO,Na), + Hg + (HgCl)».

Potassium iodide solution converts it into mercurous iodide and
potassium sulphite. Then, as in any case would happen, concentrated
solution of potassium iodide resolves this mercurous iodide into
mercury, and mercuric iodide dissolving as double iodide; but the
presence of the potassium sulphite does not appear to affect this
decomposition of the mercurous iodide. As stated in the section on
mercuric sodium sulphite, alkali sulphites do not act on mercurous
iodide as they do on mercurous chloride and mercuric iodide.

Mercurie chloride solution reacts with hypomercurosic sulphite to
form mercurous chloride and sulphuric acid :—

Hz,(SO;)2 + 20H, + 4HgCl, = 2S0,H, + 4(HgCl).,

but its action takes some time to complete, and the mercurous chloride

570 DIVERS AND SHIMIDZU : MERCURY SULPHITES,

is at first very grey. When the mercuric chloride is not in excess,
the filtered solution of sulphuric acid is free from both mercury and
chlorine, but when it is added quickly in excess and at once filtered,
a solution is obtained which continues to deposit crystalline mer-
curous chloride for some time. The reaction is, therefore, similar
to that between mercurosic sulphite and mercuric chloride. First,
mercuric sulphite (or chloride sulphite) and hypomercurous chloride
(or mercury and mercurous chloride) are formed :—

Hg(SO,).Hg, + HgCl, = Hg(SO,),Hg + Hg,Cl,(?Hg + Hg,Ch) ;

then mercuric sulphite and water become mercurosic sulphite and
sulphuric acid; the mercurosic sulphite decomposes with mercuric
chloride, as stated in describing that sulphite; and so the changes
follow on until no more sulphite remains. The hypomercurous
chloride (or mercury) slowly unites with mercuric chloride to form
mercurous chloride, according to the equation given in this section in
the paragraph on the action of hydrochloric acid.

Mercuric oxide (precipitated) has no immediate action on moist
hypomercurosic sulphite. With the oxide not in excess, the mixture
slowly changes to mercurous sulphate and mercury, a result appa-
rently not due to any direct reaction between the two substances.
With the oxide in excess, we once got a bright orange- brown mercury
oxysulphite; this has not been fully examined. It is at once de-
composed by hydrochloric acid in the usual way, but is insoluble in
dilute nitric acid; this affords a ready means of removing the excess
of mercuric oxide. A little mercurous sulphate was present with it.
Mercurie nitrate solution rapidly converts it into mercurosic sulphite,
becoming itself changed to mercurous nitrate :—

Hg(SO,).Hg, + Hg(NO;), = Hg(SO;)sHg, + (HgNOs)».

As mercuric nitrate readily takes up mercury, the above equation,
showing the transfer of one of the three atoms of the hypomercurous
radicle to the mercuric nitrate, presents no difficulty. The mercurosic
sulphite obtained in this way is always of a light buff colour, instead
of white. This colour is characteristic of mercurosic sulphite not
obtained by hydrolytic decomposition; for the reaction between
mercuric sodium sulphite and mercurous nitrate or sulphate yields
it also of this colour. Formed by hydrolytic methods, it is always
exceedingly white. Mercuric sulphate solution behaves like mercuric
nitrate, but in this case the mercurous sulphate formed precipitates
and mixes with the mercurosic sulphite. Mercurous nitrate solution
and mercurous sulphate are, as might be expected, inactive. So, too,

is silver nitrate.

AND THE CONSTITUTION OF SULPHITES. 571

Formation and Preparation.—Hypomercurosic sulphite is
formed in reactions between—

(a.) Sulphurous acid, water, and mercurosic sulphite (also mercuric
oxysulphite) ;

(b.) Sulphurous acid, water, and mercurous sulphate, or solution of
mercurous nitrate ;

(c.) Solution of sodium sulphite and mercuric oxysulphite ;

(d.) Solution of sodiwm sulphite and mercurosic sulphite ;

(e.) Solution of sodium sulphite and mercurous sulphate, or solution
of mercurous nitrate or mercurous chloride ;

(f.) Silver sulphite and solution of mercurous nitrate or mercurous
sulphate in water.

a. Mereurosic sulphite, whem treated with sulphurons acid, yields
two-thirds of its mereury as hypemercurosic sulphite, the rest dissolv-
ing as mercuric acid sulphite. Since mercuric oxysulphite with
sulphurous acid yields mercurosic sulphite, it also yields hypomer-
eurosic sulphite. The sulphurous acid must be added to these
sulphites gradually and not in excess, as otherwise it will decompose
the hypomercurosic sulphite.. The blackening of mercurosic sulphite
when heated appears to be a case of the reaction between sulphurous
acid and mercurosic sulphite:.

b. Mercurous sulphate and nitrate are at once decomposed by
sulphurous acid, even in presence of dilute sulphuric or nitric acid,
into hypomercurosic sulphite and their respective acids. As the
mercurous radiele, from the nitrate, cannot hold the non-oxylic
relation to sulphuryl, half its mercury goes to form the hypomer-
curous radicle on the oxylic side :—

2(H-SO,H) + 2(HgNO,), = Hg(SO,),Hg, + 4HNOs.

As already mentioned, hypomercurosic sulphite is destroyed, as well
as formed, by sulphurous acid, but its destruction takes place only
when the solution contains an inufficient quantity of either sulphuric
or nitric acid. In preparing the sulphite by this method, the sulphu-
rous acid solution should be mixed with a little sulphuric acid before
pouring it into the mercurous nitrate solution or on to the sulphate.
Mercurous sulphate and sulphurous acid mixed with a little sulphuric
acid, are the best source of hypomercurosic sulphite. The sulphate
should not be dense and crystallised, but flocculent, as obtained by
precipitation ; in presence of sulphuric acid, the sulphurous acid may
be added to it in excess. If concentrated sulphurous acid is added
gradually and with stirring to the mercurous sulphate in the state of
paste, the sulphuric acid can be dispensed with, but as soon as the
sulphurous acid is in slight excess, further addition of it must be

FO AS ke Oe ee Ee

iP oe

Messer Sere Fr APT

572 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

stopped; it liberates enough sulphuric acid to protect ihe hypomer-
curosic sulphite from its further action, provided an excess of it does
not too much dilute this sulphuric acid; it is safer and simpler, how-
ever, to add some sulphuric acid to the sulphurous acid. It is better
not to attempt to purify the sulphite by much washing, as it begins
at once to suffer hydrolysis by treatment with water; draining in a
thin layer on a good porous tile is the best way, and leaves mere
traces of sulphuric acid. The preparation must not be grey, even
when dry, but almost black, and in the moist state must not be clotted ;
clotting and grey colour are proof of hydrolysis having proceeded to
some extent.

c. The formation of hypomercurosic sulphite from mercuric oxysul-
phite by sodium sulphite is somewhat obscure in nature, and has
already been described. The amount of hypomercurosic sulphite
produced is proportionately small, but is certainly not dependent on
the presence of mercurous sulphate before addition of the sodium
sulphite.

d. The conversion of mercurosic sulphite to hypomercurosic sulphite
by sodium sulphite is quite similar to that which happens when
sulphurous acid is used. The sodium sulphite must not be added in
excess.

e. Sodium sulphite added to either moist mercurous sulphate or a
solution of mercurous nitrate, produces its effect by double decom-
position, but as the mercurous radicle cannot hold the non-oxylic
position in a sulphite, half its mercury is left to unite with the
oxylic mercurous radicle, and convert it into the hypomercurous
radicle :—

2Na(SO,0Na) + 2(HgNO,), = Hg(SO,0),Hg, + 4NaNO,.

The sodium sulphate solution should be dilute and be added gradually,
but not in excess, as it then rapidly attacks the hypomercurosic
sulphite. It is difficult, however, to avoid having a temporary
excess of the sodium sulphite, particularly when working with
mercurous sulphate; and the process is, therefore, very inferior to
that in which sulphurous acid is used, if a nearly pure sulphite is
desired. In the case of mercurous chloride, it is hardly possible to go
nearer to success than seeing the blackening effect of the sulphide
upon the chloride.

f. Silver sulphite can be used effectively as a means of producing
hypomercurosic sulphite from mercurous nitrate and even from mer-
curous sulphate. The insolubility of silver sulphite is so great that
it can remain for a few moments in the solution of mercurous nitrate
before beginning to lose its whiteness; when started, however, the
change proceeds rapidly and soon becomes complete if the mercurous

Se ht of Sh at a ttle

AND THE CONSTITUTION OF SULPHITES. 573

nitrate is in excess. In working with mercurous sulphate, the two
salts require intimate trituration with water for some time. The
blackening that occurs shows how fully they act on each other, but
the hypomercurosic sulphite and silver sulphate can, of course, not be
separated. The reaction between mercurosic sulphite and silver
sodium sulphite has already been described.

Chemical Composition.—Although hypomercurosic sulphite is
produced by the reaction between other sulphites and mercurous salts,
its behaviour with reagents shows that it is not a mercurous salt, for
it always yields a mercuric compound, as well as mercury and a mer-
curous compound. Its reactions with hydrochloric acid and with
mercuric chloride seem to leave its constitution as a mercuric hypo-
mercurous salt beyond doubt. The mercuric chloride and mercury
formed in this case, if well shaken together, combine almost completely
and become mercurous chloride.

That in its formation from mercurous nitrate or sulphate, the
mercury of two mercurous radicles becomes distributed as a mercuric
and a hypomercurous radicle, is in agreement with what takes place
on adding excess of sodium or hydrogen sulphate, as then mercuric
sodium or hydrogen sulphite and mercury are the products. Evidence
of the existence of other hypomercurous compounds is not altogether
wanting. Mercury shaken with mercuric chloride solution yields a
nearly black precipitate, which is slowly converted into white mer-
curous chloride by contact with more mercuric chloride. Finely
divided mercury is, of course, far removed from black in colour, but
if, in spite of this, it should be contended that the black substance
may be mercury only, it will be enough, perhaps, to point out for
reply that Vogel mistook hypomercurosic sulphite for mercury in the
state of a black-grey powder (Gmelin’s Handbook, 6), when he saw it
produced by the action of sulphurous acid on mercurous nitrate.

The reactions of hypomercurosic sulphite are clearly those of a
sulphite. Heated in a tube it yields water. Its insolubility in dilnte
nitric acid at once distinguishes it from mercurous oxide. Its colour
distinguishes it from other mercury sulphites, and also from grey
lustrous mercury, into which it so quickly changes, under conditions
similar to those in which it is formed. From mercury it is also dis-
tinguished in being flocculent and voluminous. Its behaviour, when
heated, either wet or dry, has already been described as evidence of
its quantitative composition.

Dissolved in a mixture of bromine and hydrochloric acid, and pre-
cipitated by the usual methods, it has given us the quantities of
mercury and sulphur tabulated below. Water has not been directly
determined, because sulphur dioxide is also evolved in the desiccator.

574 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

The weights taken for analysis were 0°8073 and 1°2510 grams of

different preparations.
I II. Hg,(SO;)2,0H3.

Mercury 8155 81-95 8120
Sulphur 6°50 6°44 6°54

Oxygen a= — 9°82
— — 1°84

100°00

General Constitution of the Mercury Sulphites.

The reactions, formation, and composition of the mercury sulphites
seem to establish the five propositions which follow concerning their
general constitution.

1. In mercury sulphites, the relation of basic to acid radicle is of
two kinds; it is half like that of the basic element in oxylic salts
(such as sulphates and nitrates), and half like that of the basic
element in salts of the halogens, including cyanogen.

2. The mercuric radicle, Hg, takes up and preserves the non-oxylic
relation to sulphuryl, with an energy to be compared to that it exerts
in its cyanide and iodide. This is the central conception in the
chemistry of the mercury sulphites, for without it all peculiarities
stand unexplained.

3. The mercurous radicle, Hg», never holds non-oxylic relation with
sulphuryl, and when presented in chemical reactions to a sulphite,
resolves itself into the mercuric radicle which goes to the sulphuryl,
and mercury which becomes free or changed into the hypomercurous
radicle, Hg;.

4, The mercuric radicle, in presence of water, can hold the oxylic
relation to sulphuryl in such an imperfect way only, that two-thirds
of it are in combination with simple oxygen radicles. But two such
compounds are known, both of which have exclusively the mercuric
radicle for base, namely, mercuric oxysulphite and mercuric oxy-
sulphate :—

Oxysulphite. Oxysulphate.

SO,,0HgO OHg0-SO,0Hg0. 1,
Hg<go0,.0Hg0> 48 He<oHg0-80,,0Hg0> Hs

The oxysulphite presents itself as a combination at once stable and
unstable; for while its formation actually involves the liberation of
nitric acid from a nitrate, as well as its insolubility in this acid when
dilute, yet it is afterwards exceedingly liable to decomposition. Its
formation in circumstances, the like of which would prevent that of
any other basic salt, seems due to the energy with which the mercuric

AND THE CONSTITUTION OF SULPHITES. 575

radicle and sulphuryl enter into non-oxylic union, the sulphuryl
being then compelled to take up other mercuric radicles on its oxylic
side, in order to complete the salt. It is the feeble hold of sulphuryl
on the mercuric radicles in oxylic union with it, which then, in con-
junction with the readiness of the sulphites of mercury and of silver
to undergo hydrolysis, gives to this sulphite its great instability.

5. The mercurous radicle can enter into oxylic union with sul-
phuryl, but only when this is exclusively united with mercury as base.
Mercuroso-mercuric sulphite and mercurous sulphate are, therefore,
the only mercurous salts, no double mercurous sulphates being
known :—

Mercurosic sulpbite. Mercurous sulphate.

SO.°0 0-S0,°0

Hg<sgo,-0> He Hg<o.g0,-0> He

The Nature of Reduction by Means of Sulphurous Acid.

In describing the formation and changes of mercury sulphites, most
of what we wish to say about the reduction of mercuric to mercurous
salt, and of the latter to metal by sulphurous acid, has already in a
manner been said. The subject is, however, of such interest and im-
portance, and the nature and course of the reduction has hitherto
been so superficially or so erroneously apprehended, that a section of
this paper may well be devoted to giving a connected statement of
what we have made out about it.

When sulphurous acid or a sulphite effects the reduction of mer-
cury or silver, or reduces ferric to ferrous salts, thereby becoming
sulphuric acid or sulphate, it is not really the case that the sulphurous
acid is, as such or directly, oxidised by the metallic salts, either imme-
diately, or, in the case of haloid salts, through the intervention of
water. As an expression of the result of heating sulphurous acid
with mercuric chloride, the following equation is of course true, and
for its purpose is free from objection :—

But it is wrong if the implication is intended that the sulphurous
acid and the mercuric chloride make a combined attack upon water,
and share its elements between them, the truth being that they first
attack each other (see the section on “ Mercuric Hydrogen Sulphite’’).

In most cases certainly, and probably in all, the first step in the
reduction is the formation of a sulphite of the metal with its com-
bining value undiminished. Mercuric salts yield mercuric hydrogen
sulphite, mercuric sodium sulphite or other double salt, mercuric oxy-
sulphite, or mercuric sulphite combined with chloride. According to

576 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

Berzelius, mercuric selenite and sulphurous acid become mercurous
selenite and sulphuric acid, but this is a mistake, the sulphurous acid
displaces the selenious acid, and the mercuric sulphite suffers hydro-
lysis into mercurosic sulphite and sulphuric acid. With excess of
sulphurous acid, the selenium is slowly precipitated from solution and
colours the mercurosic sulphite, thus making it look a reducing
selenite and liable to cause a mistake: mercurous selenite is not
formed. Mercurous sulphite is unknown, but mercurous salts yield
hypomercurosic sulphite metameric with it (see the previous section).
Silver nitrate, sulphate, chloride, or oxide, treated with sodium
sulphite become either silver sulphite or silver sodium sulphite, and
the nitrate and sulphate become sulphite when treated with sulphurous
acid. The reddening of ferric solutions on the addition of sulphurous
acid as well as of sodium sulphite, precedes reduction, and this indi-
cates the formation of ferric sulphite, for a similarly coloured solution
is obtained by dissolving ferric hydroxide in sulphurous acid, and
this has the properties of a solution of ferric sulphite. Dilute sul-
phuric acid or nitric acid does not prevent the formation of sulphites ;
and when from the presence of one of these acids in greater concen-
tration this formation is prevented, so also is reduction. Sulphurous
acid has scarcely any action on mercurous or silver chloride, perhaps
none if sulphuric acid and other impurities are altogether absent; so
that here, too, absence of reduction goes with non-formation of sul-
phite. It does reduce mercuric chloride, and here apparently also
forms sulphite (see section on ‘‘ Mercuric Hydrogen Sulphite”’).

A sulphite being formed, metal becomes reduced from it in one of
two ways. Mercurosic and hypomercurosic sulphites treated with
sulphurous acid or sodium sulphite are resolved into free mercury
and mercuric sulphite, dissolving as double salt of hydrogen or
sodium. The consequence is remarkable, for it follows from this that
mercury may become metal without oxidation of any sulphurous acid
or sulphite radicle whatever ; thus (ignoring the intermediate forma-
tion of hypomercurosic sulphite) the reaction between mercurous
chloride (or nitrate) and sodium sulphite, and that between mercurous
nitrate and sulphurous acid are expressed by the equations—

(HgCl), + 2Na(SO,Na) = Hg + Hg(SO,Na), + 2NaCl

Such reactions can hardly be regarded as cases of reduction, half
the metal drops out in the interaction of the substances, just as it
does when potassium iodide acts on mercurous iodide. They might,
with as little impropriety, be treated as cases of owidation by a
sulphite, since a mercuric is got from a mercurous salt.

The other way in which reduction of metal is brought about is by

AND THE CONSTITUTION OF SULPHITES.

hydrolysis, or an equivalent change. Water converts the sulphite
into metal and sulphuric acid. The hydrolysis proceeds in stages,
usually apparent :—

6Hg.(SOs). + 40H, = 4Hg;(SO;),. + 4S0,H,
4Hg;(SO3;). + 20H, = 3Hg.(SOs;). + 2S0,H,
3Hg.(SO;)2 + 60H, = 12Hg + 6S0,H2.

But this is unimportant here, since the stages are alike in the
nature of the reduction, and the reactions of the several sulphites
may be said to prove that mercury each time becomes free, but then
unites with unchanged salt so as to build up the mercurous and
hypomercurous radicles. Ferric sulphite apparently does not suffer
hydrolysis, but still it becomes reduced to ferrous sulphite by the iron
losing one-third of its hold upon the sulphite radicle. There is
nearly the same separation {between metal and sulphite radicle as
occurs when mercuric sulphite becomes mercurosic sulphite by hydro-
lysis, but the acid radicle is differently disposed of. According to
Gélis, the red solution of ferric hydroxide in sulphurous acid is a
solution of ferric sulphite, and slowly suffers, in absence of other
acid or of air, conversion into ferrous sulphite and ferrous hypo-
sulphate.

But ferrous hyposulphate is equivalent to ferrous sulphite com-
bined with sulphuric oxide, and when boiled with an acid it yields
sulphate ; in this way the ferric sulphite becomes ferrous sulphite and
sulphuric acid.

Hydrolysis therefore is equivalent to oxidation of sulphurous acid
to sulphuric acid by reduction of mercury, silver, or ferric compounds ;
bat in nature it is the displacement of metal by the hydrogen and
hydroxyl of water, in which process hydrogen takes the place of the
metal, and oxygen converts its half haloid into wholly oxylic union
with the sulphuryl. It will be seen that reduction by hydrolysis of
the sulphites is a process the reverse of that of the separation of metal
by the action of sodium or hydrogen sulphite. In the latter, mercu-
rous becomes mercuric salt by part of the mercury dropping out of
combination with the sulphite radicle; in the former, when part only
of the sulphite radicle is separated from the mercury of the water,
mercuric becomes partly mercurous or hypomercurous salt.

The Oxidation and Hydrolysis of Sulphites as Evidence of their
Constitution.

Mercury and silver sulphites are not liable to atmospheric oxidation.
In this respect, they resemble organic sulphites (sulphonates), and
differ from most inorganic ones. There seems no reason why they

578 DIVERS AND SHIM.DZU: MERCURY SULPHITES,

should be so if sulphites are dioxylic salts of thionyl, (AgO),.SO; but
when sulphites are regarded as having a hemi-haloid constitution,
the explanation at once suggests itself.

In forming a sulphite with sulphur dioxide, a basic oxide becomes
half deoxidised, just as it does when it unites with chlorine, except
that with this it gives a haloid salt and an oxylic salt, whereas with
sulphur dioxide it gives but one salt, hemi-haloid, hemi-oxylic, because
here, simply, the acid radicle is bivalent. Thus conceived, a sulphite
possessing, as it does, half-deoxidised metal, is liable to direct oxida-
tion according as its metal is readily oxidisable or not. The ready
hydrolysis of the sulphites not liable to oxidation serves to show that
sulphuryl prefers dioxylic union with metallic radicles to the half-
haloid union with them it has in sulphites; it is therefore ready in all
sulphites for union with more oxygen, but when the sulphite is of
mercury or silver it cannot directly* take up oxygen because of the
indifference of the metal.

Chlorides, to which in illustration reference has just been made,
do not oxidise, because their haloid radicle has a characteristic prefe-
rence for direct or non-oxylic union with metals; their basic radicle,
however, does in some cases oxidise at elevated temperatures, their
chlorine then becoming free. (This aspect of varying oxidisability of
haloid salts is further noticed in the section on the “ Relation of
Selenites to Sulphites.”’)

On the assumption that sulphites are thionyl compounds, there is
no way of accounting for the hydrolysis of some of them. Water
would have to be decomposed by the sulphite into its elements, its
oxygen going to the thionyl, and its hydrogen taking up oxygen
again (!) by displacing mercury or silver; or the sulphite would have
first to be changed by hydrolysis into basic oxide and sulphurous acid
or oxide, and then the latter act reducingly on the basic oxide.
Neither of these interpretations of the nature of the change is pro-
bable. It will therefore be sufficient to point out again, in connection
with the second, that on trial sulphurous acid does not become
oxidised by mercury oxide, but combines with it to form a mercuric
sulphite. Hydrolysis may be regarded as a test of the constitution of
an oxygenous salt; if it yields acid and basic oxide, or hydroxide, the

* “ Directly,” that is, as regards the argument in the text, although Traube’s

view of atmospheric oxidation is most, likely true here, as in other cases :
Na Na
20,8 | + OH; + 0, = Nao, } + H,0,.

In the report of the Inventions Exhibition by the editor of the Journal of the
Society of Chemical Industry, Messrs. Boake and Co. are suid to have exhibited
potassium metasulphite (pyrosulphite) having the advantage of not being liable to
oxidation. This property will of course be lost in presence of moisture.

AND THE CONSTITUTION OF SULPHITES. 579

salt is wholly oxylic; if it yields acid and metal (hydrogen or hydro-
carbon), the salt is partly haloid.

Sulphites resemble Chlorides, Iodides, and Cyanides in Properties.

Mercury and silver sulphites possess certain properties which are of
value as indications of the partly haloid character of their acid radicle,
because they belong more or less fully to haloid as distinguished from
oxylic salts. Such properties are, that of decomposing in ways
equivalent to a separation of the metal from the salt radicle; relative
insolubility in acids; and that of forming double salts when possible.

Silver and Mercury Sulphites Decompose much in the same way as
some Halvid Salts—Mercuric sulphite is unlike any purely oxylic salt,
and resembles many haloid salts, in yielding up a third or a half of
its chlorous radicle, and thus becoming mercurosic or hypomercurosic
sulphite. Thus, manganese tetrachloride becomes a lower chloride
by giving up a quarter or a half of its chlorine; cupric and ferric
iodides become cuprous and ferrous by setting free some of their
iodine ; cupric cyanide becomes cuprous by parting with cyanogen,
and so in other cases. Again, much as these sulphites become metal
and sulphuric oxide in hot water, gold chlorides by a moderate heat
become metal and chlorine. The differentiation of a sulphite from a
purely oxylic salt is here perfect, and although its resemblance to a
simple haloid salt is not complete, the conditions of its decomposition
being unlike those of a haloid in including the presence and activity
of water, there is good reason for this in the consideration that the
sulphite is half oxylic in constitution, and consequently requires the
assistance of water in decomposing.

The Comparative Insolubility of Silver and Mercury Sulphites in
Acids.—Befvre trying to show the significance of this fact, it may be
well to point out, more distinctly than is done in Gmelin’s or other
handbooks, the extent of the insolubility of silver sulphite.

Silver nitrate solution is copiously precipitated by sulphurous acid.
The reaction is immediate, and so certain that silver nitrate solution
proves as sensitive to either gaseous or dissolved sulphur dioxide as
lime-water is to carbon dioxide, whilst the brilliant whiteness of the
silver sulphite makes the effect much more visible than the lime-
water effect, and thus well fitted for class demonstrations, and as a
test for sulphur dioxide. Except when the silver nitrate is in con-
centrated solution, the precipitation of silver is almost complete.
Nitric acid is, of course, liberated in this reaction, and so insoluble is
silver sulphite in this acid as to be hardly affected until the acid is
strong enough to oxidise it and become itself reduced to nitrous fumes.
Excess of sulphurous acid has no action on the precipitate, for a time
at any rate, in presence of the nitric acid which has been set free.

580 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

Silver sulphate is quickly converted by sulphurous acid into silver
sulphite and sulphuric acid. On the other hand, sufficiently concen-
trated or hot dilute sulphuric acid decomposes the sulphite, reversing
the change.

The great insolubility of mercurosic and, above all, hypomercu-
rosic sulphite, in nitric acid, and even sulphuric acid, is noticed in
other sections of this paper; and so is the behaviour of mercuric
oxysulphite towards acids, but this it will be well to describe again
in the present connection because of its significance. Although a
basic salt it, like the other sulphites, is insoluble in rather dilute
acids, as might indeed be expected from the fact of its formation
being attended with liberation of nitric acid. When it does dissolve
in stronger acid, it does so with little or no liberation of sulphur
dioxide ; the solution, indeed, changes rapidly, and mercurous sulphate
is the result, but if, without delay, hydrochloric acid is added to the
solution, very little mercurous chloride is precipitated, and abundance
of sulphur dioxide is given off—effects which show that the sulphite
dissolves as such. Hence it will be seen that the insolubility of
mercury and silver sulphites is not, in some way or other, a cause of
their not being decomposed by acids, but a result rather of their
resisting decomposition.

To come now to the interpretation of this resistance of silver and
mercury sulphites to the action of oxylic acids, as evidence that
sulphites have the character of haloid salts. In the first place, if
these sulphites were simply oxylic salts they would be the only oxylic
salts of silver and mercury insoluble in dilute nitric acid, the sul-
phates in part excepted; while as haloid salts they are far from excep-
tional in this respect. Stronger acids decompose them, and so does
strong sulphuric acid decompose, to some extent, mercury haloid salts.
The action of nitric acid on silver sulphite is the same as on silver
cyanide (except that it is more easily exerted than on the latter), these
salts being unchanged by cold dilute acid, and decomposed with
deoxidation of the nitric acid when this is hot and strong.

In the second place, if sulphites had to be ranked simply as oxylic
salts, it would be incomprehensible that silver and mercury sulphites
should be insoluble in dilute sulphuric acid, and be freely formed from
the sulphates by sulphurous acid. For sulphurous acid not only proves
to be much weaker than sulphuric acid in salts of other metals, but as
the elements of the two acids are the same, the acid with more oxygen
should be the stronger acid. As haloid salts, on the other hand,
silver and mercury sulphites are not peculiar, since, as every one
knows, not only do the strong haloid acids decompose silver and
mercury sulphates, but so also does hydrogen cyanide, feeble as it
is towards alkalis. Sulphites, as haloid salts, have an essentially

AND THE CONSTITUTION OF SULPHITES. 581

different constitution from that of sulphates, half their metal being
directly united to the sulphur of sulphuryl; so that mercury and
silver sulphites form and exist in the presence of sulphuric acid, for
much the same reason that mercury mercaptide proves to be a very
stable body, although mercury ethoxide cannot even exist.

The Stability of the Double Sulphites resembles that of the Double
Cyanides, Iodides, and Chlorides. —A reference to the section on
“Mercurie Sodium Sulphite ” will show how this salt resembles the
corresponding iodide and cyanide, in the circumstances under which it
forms, and in its behaviour to reagents. Silver sodium sulphite is but
little behind the mercury salt in this respect. Sodium hydroxide and
chloride do not precipitate the silver from the double salt, whilst silver
oxide, chloride, and other salts of silver dissolve in sodium sulphite.*
The double sulphites, like the double cyanides, are neutral to litmus,
although alkali cyanides and sulphites are strongly alkaline.

Plainly in the matter of forming double salts, sulphates resemble
haloid salts. Many oxylic salts do indeed form double salts, but not
under such conditions as those in which the mercury and silver double
sulphites can form. These wholly oxylic double salts are able to
resist little more than the action of water, and appear to be held
together mainly by forces of crystallisation only. Such salts of
mercury and silver are but little known, and are decomposable even
by water.

In the same way that sulphites resemble haloid salts they are like
thiosulphates, in which also the metal is half oxylic—AgS-SO,-ONa.
Hyposulphates, so closely related to sulphites in origin, show no
relation to haloid salts; they are wholly oxylic—(SO,0Na).

The Constitution of Sulphites a Clue to that of Cyanides and other
Halides.

The simple halogens, and cyanogen, like the radicle of the
sulphites, form double salts of a little oxidisable metal and a highly
oxidisable metal, which are more stable than their salts, formed
exclusively of one or the other kind of metal. It seems clear that
this two-sided character of the sulphite radicle in its combination with
metals is due to its constitution, through which it acts half as an
oxylic radicle. But, this being the case, the halogens should owe
their two-sided character as acid radicles also to peculiarities of con-
stitution. Cyanogen is a compound radicle, and admits of being dealt
with from this point of view. Chemists are not decided as to whether

* This was pointed out in a previous paper by one of us last year. Captain
Abney has recently recommended the use in photography of sodium sulphite, in
place of thiosulphate.

VOL. XLIX. 2k

582 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

cyanides are carbamines or nitriles ; more commonly they are repre-
sented as the latter, but there is much to be said in favour of their
being carbamines, as pointed out by one of us in a previous paper
(Trans., 1885, 227). Now by giving them a constitution half-carbamine,
half-nitrile, an explanation is afforded of their remarkable tendency
to be double salts. Writing them thus—

Potassium cyanide. Silver potassium cyanide.

KN:C:N:CK AgN:C:N:CK,

with half their metal united to nitrogen and half to carbon, it would
seem certain that silver or mercury, from its affinity for nitrogen,
would form a more stable combination than potassium on the nitrogen
side of the cyanogen, and will not do so on the carbon side.

The stability of mercury potassium iodide is no less remarkable
than that of the cyanide or sulphite, although in it the chlorous
radicle is a simple halogen. When hydrogen iodide is put in contact
with a mixture of potassium hycroxide and mercuric oxide, it does
not take one and leave the other, but satisfies itself half with one and
half with the other. It would, therefore, seem that in Hg(I.K),, the
group l1,(or I,) has a structure, constitution, or internal relation of
parts, by which it presents to basic radicles two forms of combining
power. Why should it not? Turning to a kindred matter, well
before the attention of chemists, we will reply by another question.
Which, we would ask, is more probable: that the resemblance of the
periodic series of the elements to the homologous and isologous series
of organic compounds is a chance one; or that it is due to the
elements being themselves complexes ?

Nore on THE Constitution oF Futminates, By Epwarp Divers.—
It will be seen that the matter of the section to which this note is
appended, bears on the question as to the nature of the fulminates,
since they show so marked a difference in their two units of basicity.
I hope yet to publish again on their formation and constitution, as
some work now going on in my laboratory, on mercury and nitric
acid, promises to throw light on the matter; but in the present con-
nection I wish to point out a modification of the formula I have
proposed for a fulminate, which, without much altering its general
significance, would make it fulfil two desirable ends. Instead of

AgOCNONCAg, I would write AgOCNONC, that is, remove Ag from
Ag

C to-an N, disconnect the Cs, and read the Cs and Ns with valencies

the same as they have in carbamines, either C” and N’” or C!’ and

N’. Thus, first, the advantage is gained of representing the carbons

disconnected, in accordance with the evidence afforded by the action of

AND THE CONSTITUTION OF SULPHITES. 583

chlorine and iodine on the fulminates (Trans., 1885, 78), that ful-
minates do not yield dicarbon compounds ; and, secondly, of having
the silver atom which is not displaceable by potassium standing in
union with nitrogen instead of carbon. I have always felt the dis-
connection of the silver from the nitrogen to be a weak point in my
formula, but, as stated in my first paper on the subject (Trans., 1884,
22), the supposed formation of fulminates from nitrous acid and
alcohol make this disconnection difficult to allow. This formation
having since then been disproved by Mr. Kawakita and myself (loc.
cit., 27), the difficulty it caused is removed ; and now the study of
mercury sulphites further inclines me to assign to the mercury or
silver in a fulminate a direct connection with nitrogen, as in the
modified formula given above.

Sulphur, Selenium, and Tellurium Sulphowides are Sulphites.

In this Journal (Trans., 1885, 218, and in earlier papers, in conjunc-
tion with Mr. Shimosé), one of us has given to the compounds which
sulphuric oxide forms with sulphur, selenium, and tellurium, formule

like this: 0,8ZO\Te—TeZSO0.,, in which the element is repre-
sented as having gone half into direct union, half into oxylic union
with the sulphuryl. Now, doing this makes them sulphites in consti-
tution, and such they must apparently be taken to be.

They are hydrolysed by water partly into sulphurous acid and what
is equivalent to a base :—

Se.(SO;).2 a 2H,0 = (Se a SeO,) +> H.SO,.

They are also partly hydrolysed, like silver sulphite, into metal and
sulphuric acid :—

Te.(SO;). + 2H,O0 = 2Te + 2H,SO,.

They stand in the same relation to the sulphites of strong metals
as the corresponding chlorides do to metal chlorides. They resemble
salts of the halogens in some of their relations. They are not only
formed by the direct union of their quasi-metal with the suiphur
trioxide, but can to a certain extent be decomposed again into these by
heat in a vacuum; thus, the selenium compound (Trans., 1885, 203),
decomposes much as platinous chloride does :—

Se.(SOs)2 — 2Se _ (SOs). Pt.Cl, => 2Pt + 2Cl,.
The selenium compound, like the chlorides of selenium, dissolves in
fuming sulphuric acid without decomposition. When thus dissolved,
it is decomposed, like a sulphite. by hydrochloric acid :— '

Se.(SO;). + 2Cl1H = Cl,Se,(SO;) + SO,H,,
2Rr2

584 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

and this compound, not yet isolated (loc. cit., 195), reacts with more
hydrochloric acid, just as the corresponding chloride sulphite of
mercury (section of this paper on “ Mercurosic Sulphite”’) does with
water :—
Cl,Se.fSO;) + CIH — Cl.Se, + CIHSO,
Cl,Hg(SO;) +(HO)H = Cl,Hg. + (HO) HSO,.
Admitting these compounds to be sulphites, a further illustration that
sulphites resemble haloid salts is seen in their formation direct from
metal and sulphuric oxide :—
2Te + (SOs). — Te.(SOs;)2.
In a previous paper by one of us, these compounds have been

treated as substitution representatives of sulphuric peroxide, SQ,,
thus :—

Lysl Ls]
O O O O

and such they remain when regarded also as sulphites; for just as
hypochlorous oxide may be regarded as chlorine hypochlorite, so may
sulphuric peroxide (if SO, expresses its composition), be regarded as

oxygen sulphite, oZ°\so., or the compound CIHSO, as chlorine
hydrogen sulphite, HO } goc,, Which in reactions it proves to be.
Cl ws

Nore sy E. Divers on Sutpnuric Oxipz.—In the paper on “The
Constitution of some Non-saturated Oxygenous Salts ” I was led to
discuss the constitution of sulphuric oxide. I pointed out the strong
probability there is of this oxide being sulphuryl sulphate, and in
@ more recent paper, in conjunction with Mr. Shimidzu, I gave
further experimental evidence in favour of this view. I would now
make good an omission in not calling attention to a synthesis of
sulphuric oxide, effected many years ago by Odling and Abel (Chem.
Soc. J., 7, 2), which leads to the adoption of the same view.
Sulphuryl bromide and silver sulphate heated together yield silver
bromide and sulphuric oxide :—

In using the name sulphuryl sulphate, it is not suggested that there
is any difference in the two sulphur-atoms, sulphuryl sulphate being
neither more nor less than sulphuryl oxide doubled :—[0,S : O,: SO].

The Relation of Selenites to Sulphites.

With the exception of a few unstable compounds, all sulphites have
the hemihaloid or sulphonic constitution. It may be otherwise with

AND THE CONSTITUTION OF SULPHITES. 585

selenites in consequence of the more basylous character, yet, on the
whole, much weaker combining power, of selenium; and dioxylic
selenites may be the common form.

Ethyl selenite is dioxylic and easily prepared ; ethylselenonic acid
or ether is unknown. The existence of methylselenonic acid is not
certain, for Woéhler and Dean’s results are not conclusive on this
point.

Metallic selenites are generally considered to have a family resem-
blance to sulphites ; or, perhaps, it would be more correct to say that
they are assumed to have it. If they really are like sulphites, then
they cannot but have the same constitution, but if they are not like
sulphites they may differ in constitution and be dioxylic. They do
exhibit many unlikenesses to sulphites, but these are hardly such as
would suggest that the selenites differ from the sulphites in being
dioxylic. Their differences are, rather, such as may be attributed to
selenious acid being a more condensed acid than sulphurous acid, and
to the feebler affinities of selenium. If, though hemihaloid, selenites
do not oxidise in the air like sulphites, that is only because for oxida-
tion to take place it must proceed simultaneously on the metal and on
the chlorous radicle: therefore, because of the selenium, sodium
selenite does not oxidise; and because of the silver, silver sulphite
does not; whilst sodium sulphite does oxidise because the sodium and
sulphur take the oxygen between them. (See, further, the section on
the “ Oxidation of Sulphites.”)

Among the undoubted resemblances of selenites to sulphites, are
some of those peculiarities which distinguish sulphites from oxygen-
ous salts generally, and connect them with haloid salts. Mercury and
silver selenites have a degree of insolubility in nitric acid which,
though less than that of sulphites, is still far more than sufficient to
enable free selenious acid to precipitate them from the nitrates. Mer-
curic selenite* is soluble in potassium selenite; so is mercuric oxide ;
and the solutions are not precipitable by potassium hydroxide.
Further agreement fails. Mercurous selenite is not decomposed by
potassium selenite into mercuric salt and mercury, and silver selenite
does not dissolve in potassium selenite. This is somewhat remarkable,
but is probably due to the affinities of selenium being so much weaker
than those of sulphur and to silver forming less stable compounds
than mercury with selenium, as well as with sulphur, silver sulphite
being completely decomposed by mercury nitrates. The points of
agreement, however, between selenites and sulphites leave but little
doubt that the former, like the latter, have a hemihaloid consti-
tution.

* The insoluble normal selenite. We have not succeeded in preparing Berzelius’s
soluble acid selenite.

586 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

But when silver selenite is treated with ethyl iodide, it yields.
dioxylic selenious ether, from which metallic selenites would appear
also to be dioxylic, as Michaelis and Landmann have suggested they
are, instead of being hemihaloid like metallic sulphites. It cannot be
admitted, however, that the iodide test is sufficient by itself to settle
the constitution of an oxylic salt, for it gives conflicting testimony
as to the constitution of nitrites. Only silver selenite yields the
selenious ether; with potassium selenite it is difficult to get a reaction,
and when this occurs, it leaves the selenium wholly deoxidised.
Since ethylselenonic acid cannot be prepared by the oxidation
method, it does seem allowable to suppose that in the reaction of the
silver selenite a selenylic group changes into a selenosic one, so as to
enter into oxylic union with the ethyl. To suppose this, is a way out
of the difficulty certainly much easier than it would be to consider
metallic selenites as dioxylic salts.

The Analogue of Sulphurous Acid ameng Carbon Compounds.

In the discussion of the paper by one ef us on the constitution of
some non-saturated oxygenous salts, Professor Tilden drew attention
to the modification that might be needed in the present view of the
constitution of carbonates, if that which was therein advocated for
the sulphites should be adopted. The point then raised is a very
interesting one, and has suggested to us a further ground for recog-
nising the sulphonic constitution of sulphites, and also for regarding
sulphur as only quadrivalent in sulphates as well as in these salts.

Beyond the fact that both sulphites and carbonates are easily de-
composable salts, there is nothing specially to connect the two classes
of salts. Indeed, even as regards their decomposability, the two
classes could not be more unlike than they are in their mercury and
silver members. Besides, carbonates are not oxidisable, whilst sul-
phites are readily oxidisable. It would, therefore, seem strange to
require that a similar constitution should be found for the two classes.
of salts. Rather than this, there is surely a strong presumption that
their constitution must be unlike. Nothing is more certain than that
carbonates are dioxylic salts; nothing, therefore, would seem more
likely than that sulphites should not be dioxylic.

Sulphites find their real analogue among carbon compounds, not
in carbonates, but in formates, although differing from these in
having the double atom of oxygen in their radicle, and in having the
hydrogen, which in the acid is joined to the radicle, displaceable by
metal.

Both sulphites and formates are monoxylic; both suffer oxidation
to salts of dioxylic acids; and both suffer hydrolysis with similar

AND THE CONSTITUTION OF SULPHITES. 587

effects when the basilous radicle is silver or mercury. The oxidi-
sability of formates has always been remarkable, as contrasted with
the high degree of stability of the acetates (and higher homologues),
in presence of powerful oxidising agents. Now the difference in this
respect between formates and acetates is due to the fact that the last
remaining hydrogen-atom of methane (aldehydic hydrogen) is readily
oxidisable, whilst methyl in place of it is not oxidisable. But this is
precisely analogous to what was said in the former paper as to the
reason why most metallic sulphites are oxidisable and organic sulphites
are not. Methylsulphurous acid and acetic acid—

HO>80, po>Co,

do not oxidise, whereas sulphurous acid and formic acid do oxidise.

HO>SO: + 0 = (HO):80,, 74>CO + 0 = (HO).CO.

In this oxidation, the carbon does not change in valency ; therefore,
there is no reason to consider the sulphur as doing so. There is good
evidence of sulphur being quadrivalent in sulphites, and therefore
also in sulphates. If the sulphur is quadrivalent, the oxygen of
sulphuryl forms a bivalent double atom ; and that it does so is ren-
dered still more probable by its apparent solidarity in the changes
which the sulphuryl may undergo,* and by sulphury]l functioning so
very like carbonyl.

The hydrolysis of silver sulphite affects, of course, the whole salt,
because its acid is bibasic though monoxylic, whilst the hydrolysis of
silver formate affects only half the salt; in both cases, metal is re-
placed by the hydrogen of the water :—

AgO
He

AgoO ‘
g\s0>CO + OH: = 24g + (HO),00 + HO>c0.

>SO, + OH, = 2Ag + (HO),SO,

Sulphurous acid and formic acid both readily convert mercuric
chloride to mercurous chloride. Sodium sulphite and sodium formate
are both producible by the action of sodium hydroxide on the re-
spective acid oxides—SO, and CO.

Formic acid being the analogue of sulphurous acid, carbonic acid
represents sulphuric acid; carboxylic acids sulphonic acids (a relation
which has long been recognised); and oxalic acid, hyposulphuric

* If thionyl chloride and silver nitrate do really produce in the first instance
nitrosyl chlorosulphonate, as, according to Thorpe, they appear to do (Trans., 1882,
297), an exception after all exists to the statement that oxidation to sulphuryl does
not occur in two stages (see Trans., 1885, 215).

588 DIVERS AND SHIMIDZU: MERCURY SULPHITES,

acid. Hyposulphuric acid by hydrolysis becomes sulphurous and
sulphuric acids; oxalic acid by the same kind of change becomes
formic and carbonic acids. The analogy is complete, and surely it
must be the case, therefore, that sulphites are only monoxylic salts.

Action of Phosphorus Oxychloride on Sulphites.

In a note to the paper on this subject (Trans., 1885, 207) one of us
expressed a doubt as to whether Carius had ever tried, as he was said
to have done, the action of phosphorus oxychloride on sulphites.
This was done because the results of our own experiments differed
apparently from those given on his authority, and because the reference
in Watts’s Dictionary to his paper was found to be wrong. In con-
sequence of the expression of this doubt, Professor Schorlemmer, at
a subsequent meeting of the Society, gave the correct reference to
Carius’s memoir, Annalen, 106, 330, and also some account of what
Carius had observed. The back volumes of the Annalen have since
fortunately come within our reach,* and we have now seen Carius’s
own words.

The first point important to notice is that he calls attention to the
difficulty of getting any sulphite quite dry without decomposition ; he
succeeded only with the calcium salt, and even that had to be kept
in a current of dry air for a long time at 150°. This confirms the
surmise contained in the former paper that, in our experiments the
sulphites, well-dried as they were supposed to be, must have retained
some moisture; only slight action was observed, even when the
phosphorus oxychloride was repeatedly distilled off from the sulphite
and run back upon it. It was therefore stated that “it is probable
that thoroughly dry and pure sulphites are not acted on at all by pure
phosphorus oxychloride ;” this is just the result of Carius’s more
accurate experiments. Calcium sulphite is not acted on at all by
phosphorus oxychloride at common temperatures, nor even, except
to a very slight extent, when the temperature is maintained for some
time at 120° under pressure ; but it is acted on freely at 150°.

The nature of this action is the second important point in connec-
tion with the constitution of sulphites. To make thionyl chloride—for
that was his purpose, and not the investigation of the constitution of
sulphites—Carius first took 2 mols. of phosphorus oxychloride to
3 mols. of calcium sulphite, in accordance with the equation—

3CaSO; + 2POCI, = Ca;(PO,), + 3SOCI.

Proceeding in this way he failed, not however for want of a reaction
at 150°, but because he got large quantities of sulphur dioxide (which

* A set has lately been added to the library of the Imperial University of Japan.

AND THE CONSTITUTION OF SULPHITES. 589

burst his sealed tubes). An equation such as was used in the former
paper might serve to express this generation of sulphur dioxide :—

6CaSO, + 2POCI, = Ca3(PO,). + 3CaCl, + 6S0O.,

and, so far, nothing is seen to support the view that sulphites are
thionyl compounds—dioxylic salts. As a test of their chemical con-
stitution the reaction has clearly failed, if it does not, indeed, favour
their having a sulphonic constitution.

But some thionyl chloride is also formed, which makes the matter
much less simple; and, besides, half only of the phosphorus oxy-
chloride taken is accounted for in the equation just set down. How
material the rest of the oxychloride is to the production of the thionyl
chloride, appears from the effects of using a greater proportion of
oxychloride in the experiment. Carius found, after repeated trials,
that by raising the proportion of oxychloride until equal molecular
proportions of the two substances were present, the proportion of
thionyl chloride greatly increased, and tbat of sulphur dioxide became
very small. Calcium sulphite can thus be made to yield most of its
sulphur as thionyl chloride, but only by using excess of phosphorus
oxychloride, since without this excess sulphur dioxide is the main
product. Now, this sulphur dioxide must be derived from thionyl
chloride should phosphorus oxychloride really act on calcium sulphite,
and then form thionyl chloride by this action.

For this, however, to be the true explanation of the production of
the sulphur dioxide, some of the sulphite and the oxychloride should
remain without acting on each other; because in one case the presence
of the phosphorus oxychloride must be supposed to preserve in some
way the thionyl chloride from: conversion into sulphur dioxide, or in
the other case some calcium sulphite must react with the thionyl
chloride so as to generate sulphur dioxide :—

CaSO, + SOCl, a CaCl, + 2S0,.

This is by no means an impossible reaction, but it is as yet unknown
to occur, while its doing so here would depend, as above stated, on
the improbable condition that the calcium sulphite to take part in it
is to be had as wanted even in presence of phosphorus oxychloride,
which for some reason has become inactive on it. This, indeed, seems
to remove all chance of its being what actually happens; for just in
the case where the phosphorus oxychloride is used in abundance, there
is thionyl chloride got in abundance, and nearly without sulphur
dioxide. The other view of the matter, that phosphorus oxychloride
may in some way preserve the thionyl chloride from decomposition,
does not need examining. If this is agreed to, then there is nothing
to be said for the view that thionyl chloride is the source of the
sulphur dioxide.

590 DIVERS AND SHIMIDZU: MERCURY SULPHITES.

Taking into account these three points concerning the production of
thionyl chloride from a sulphite—that a high temperature is neces-
sary; that an excess of phosphorus oxychloride must be present ; and
that sulphur dioxide is the principal and primary product when no
more than the normal quantity of the reagent is used—it cannot be
allowed that the thionyl chloride is the direct product of a reaction
between the sulphite and the oxychloride. It is not difficult to find
another origin for it.

Gustavson has shown that when phosphorus oxychloride is heated
at the temperature Carius employed, 150°, along with boron oxide,
not only is some of the boron oxide converted into chloride, but
phosphorus pentoxide is produced and unites with unchanged boron
oxide. What takes place when it is heated at 150° with calcium
sulphite appears, then, to be this: in its own form it remains without
action on the sulphite; it behaves potentially as, or else it gradually
dissociates into, phosphorus pentoxide and pentachloride; the pent-
oxide takes calcium oxide from the sulphite, thus setting free sulphur
dioxide; and this with the phosphorus pentachloride gives thionyl
chloride and phosphorus oxychloride again (Schiff’s reaction). The
following equations serve to express these reactions :—

10POCI,; = 2P,0; + 6PCI,,

6CaSO,; + 2P,0; = 2Ca;(PO,). + 6SO.,

6SO, + 6PCl; = 6SOCI, + 6POC\,.
Three-fifths of the phosphorus oxychloride begun with are thus
regained. In practice, Carius found an extra half of the oxychloride
to be excess enough to use, which is quite intelligible, since the excess
can serve over and over again. The large quantity of sulphur dioxide
which forms when these proportions are greatly altered in favour of
the calcium sulphite, is to be traced to the action of the phosphorus
pentachloride upon the excess of calcium sulphite instead of upon the
sulphur dioxide, as is shown by the following equation, intended to
take the place of the last one in the scheme of reactions just formu-
lated, on tle supposition that 15 mols. of calcium sulphite are present
in addition to the 6 mols. there set down :—

15CaSO; + 6PCl, = 3Ca,(PO,). + 6CaCl, + 9SOCI, + 680.
With the other 6SO, liberated, but now not decomposed, there are
thus obtained 12SO, to 9SOCI.. The calcium chloride here shown is
a necessary complement to the sulphur dioxide, whatever version of
the change be adopted.

In spite, then, of the production of thionyl chloride when sulphites
are strongly heated with phosphorus oxychloride, it still remains true
that the changes observed afford no support to the view that those salts
are dioxylic and decomposable according to the equation—

3CaSO, + 2POCI, = (CaO,)3;(PO),. + 3SO0Cl,.

LVI.—On the Electrolysis of Aqueous Solutions of Sulphuric Acid,
with special Reference to the Forms of Oxygen obtained,

By Hersert McLeop.

Durixe the discussion on Electrolysis at the meeting of the British
Association at Aberdeen last year, Dr. Lodge remarked that it was
well known that when Wollaston wires are used as electrodes in the
decomposition of acidulated water, a considerable quantity of the
oxygen is liberated in the form of ozone.

The following papers contain references to the production of ozone
by this process :—

Shénbein (1840), Ann. Phys. Chem., 50, 616—635. “ Beobachtun-
gen iiber den bei der electrolysation des Wassers und dem Ausstrémen
der gewéhnlichen Electricitaét aus Spitzen sich entwickelnden Geruch.”
Five or six times diluted sulphuric acid [from 23°5 to 26°9 per cent.
by weight] is best. Phosphoric acid and other solutions when elec-
trolysed produce the same odour.

De Marignac (1845), Compt. rend., 20, 808—811. “Sur la produc-
tion et Ja nature del’ozone.” The liquid used was water charged with
sulphuric acid and kept cool. The size of the electrode, the strength
of the solution, and the quantity of ozone produced are not stated.

Williamson (1845), Mem. Chem. Soc., 2,395—398. ‘Some Experi-
ments on Ozone.” Liquid used consisted of 1 volume of sulphuric
acid and 3 of water [35 per cent.]. No measurements of the quantity
of ozone are given, as the experiments were made on the assumption
that ozone was a peroxide of hydrogen.

H. Meidinger (1853), Annalen, 88, 57—81. “Ueber voltame-
trische Messungen.” The acid used had a density of 13. Small
quantities of ozone were produced and large quantities of “ peroxide
of hydrogen.”

Baumert (1853), Ann. Phys. Chem. 89, 38—55, and Phil. Mag. [4],
6, 51—63. “ Ueber eine neue Oxydationsstufe des Wasserstoffs und
ihr Verhiltniss zam Ozon.” Electrolytic oxygen obtained by elec-
trolysis of dilute sulphuric acid with large positive plate and small
negative electrode passed through solution of potassic iodide and
afterwards through sulphuric acid, and the gain of weight of the
absorption apparatus determined. In one experiment, an increase of
0°0133 in six days was obtained, and in another 0°0149 in eight days.

H. Meidinger (1854), Chem. Soc. J., 7, 251—255. ‘On the
occurrence of Ozone and Peroxide of Hydrogen in the Electrolysis of
Sulphuric Acid.” The positive electrode used was 20 mm. long and

592 MCLEOD ON THE ELECTROLYSIS OF

4mm. thick. The acid seems to have had a density of 14. 253 c.c.
of hydrogen and 40 c.c. of oxygen were evolved in 19 minutes. When
the oxygen was heated, no change of volume could be perceived, and
when the gas was acted on by solution of potassic iodide there was a
diminution of volume equal to zj5th. The acid that had been
electrolysed decomposed potassic iodide and so was supposed to
contain peroxide of hydrogen, but no determination of its quantity
was made. He thinks that if the temperature were sufficiently
lowered, pure ozone might be produced. The best acid to give
hydrogen and oxygen in proper proportions has a density of 1:1, the
negative electrode should be a wire and the positive a plate of not
too small dimensions,

Andrews (1855), Phil. Trans., 1856, 1—14. ‘‘On the Constitu-
tion and Properties of Ozone.” The acid used was made by mixing
1 volume of sulphuric acid with 7 of water [20°81 per cent.]. The
positive electrode was a bunch of fine wires, and the negative electrode
a platinum plate; the liquid was cooled by ice and water. The
electrolytic oxygen was passed through a solution of potassic iodide
acidified with hydrochloric acid, at the rate of about 750 c.c. of gas
per hour. The acid solution was found to give the same results as a
neutral solution [a statement contested subsequently by Brodie].
After passing through the potassic iodide, the gas was conducted
through drying tubes and the increase of weight determined. The
quantity of iodine liberated was also determined volumetrically.
Mean of five experiments gave 0°0041 gram of ozone for 1 litre
of gas. If ozone is O;, this number must be multiplied by 3, giving
00123 gram or 0°85 per cent. [He says, “ Pure water does not absorb
ozone.

Tyndall (1862), Phil. Trans., 1862, 84—86. In his second
memoir—“On the Absorption and Radiaton of Heat by Gaseous
Matter,” used electrolytic oxygen. Small electrodes, less than one
square inch in area, were employed, but the exact size is not stated.
The quantity of ozone was not determined, but he says, “The quantities
of ozone with which I have operated must be perfectly unmeasurable
by ordinary means.”

L. Soret (1863), Arch. des Sciences, 16, 208—215; Compt. rend., 56,
390—393. “Sur la production de l’ozone par l’électrolyse et sur la
nature de ce corps.” The electrodes were very fine wires of platinum-
iridium; when platinum is used, a black deposit is formed on the
wires which destroys some of the ozone. The acid used contained
1 volume of sulphuric acid to 5 of water [26-9 per cent.]. Ata
temperature of 5° or 6°, oxygen containing 1 per cent. of ozone was
produced (on the assumption that to one equivalent of iodine liberated
from the potassic iodide corresponds one equivalent of ozone considered

AQUEOUS SOLUTIONS OF SULPHURIC ACID. 5993

as an allotropic modification of oxygen). When the acid was cooled
with ice and salt, 2 per cent. was formed. [As with Andrews’ results,
these numbers must be multiplied by 3, so that 3 and 6 per cent. were
really obtained. ]

Brodie (1864), Chem. Soc. J., 17, 293. In a paragraph at the end
of his paper “On the Organic Peroxides theoretically considered,”
states that peroxide of hydrogen is not formed during electrolysis,
and he thinks that the peroxide of sulphuric acid, H,SO,, is present.

C. Hoffmann (1867), Ann. Phys. Chem., 132, 607—618. “ Einige
Versuche iiber die bei der Elektrolyse des Wassers auftreten-
den Mengen von Ozon und Antozon.” A negative wire used and
a positive plate, but the dimensions are not given. Gases were not
collected, but the relative oxidising action of the gas and the liquid
was determined.

Berthelot (1878), Compt. rend., 86, 71—76, and Ann. Chim. Phys.
[5], 14, 354—361. Acid of density 1°4 was used ; the positive electrode
was of fine platinum wire, sealed into a glass tube. The size
of the wire is not given. 910 c.c. = 1°250 grams of oxygen were
collected and contained 66 mgrms. of ozone [0528 per cent. ].
Active oxygen in the liquid (not hydroxyl) was 44 mgrms. This
Berthelot supposes to be present as persulphuric acid. The liquid
effervesced. Phosphoric acid gave ozone, but no oxidising liquid.
When sulphuric acid, H,SO,,H,O, was used, a small quantity of
peroxide of hydrogen was formed containing a quantity of active
oxygen, only ;),th of that present as persulphuric acid. With the
acid, H,SQ,,3H,0, only half this quantity of peroxide of hydrogen
was produced.

Schone (1873), Ber., 6, 1224—1229. “Ueber das Verhalien von
Ozon und Wasser zu einander.” The percentage of ozone by volume
in the electrolytic oxygen varied between 3°29 and 8°63.

Carius (1874), Annalen, 174, 1—30. ‘“Verhalten das Ozons
gegen Wasser und Stickstoff.” The acid used was made by mixing
1 “part” of sulphuric acid with 5 of water, and the positive
electrode consisted of a very thin platinum-iridium wire 15 mm. long.
100 c.c. of the oxygen contained quantities of ozone varying between
1°438 and 3°44 c.c.

Many of the papers just cited were consulted after the completion
of the experiments described in this paper, and it will be seen that
some of the results obtained have been anticipated by other experi-
menters.

As it appeared to be of interest to determine what quantity of
ozone is generated during the electrolysis, a series of experiments was
instituted in the hopes of throwing some light on the phenomena of
the electrolysis of dilute sulphuric acid under various conditions.

594 ELECTROLYSIS OF AQUEOUS SOLUTIONS OF SULPHURIC ACID.

In the preliminary experiments, the hydrogen which was evolved
from a pole of platinum foil was collected and measured, and the gas
from the positive pole, which was of fine wire, was passed through
a solution of potassic iodide mixed with hydrochloric acid. The
quantity of ozone was calculated on the assumption that only hydro-
gen, oxygen, and ozone were produced during the electrolysis. Sub-
sequently the oxygen was also collected, when it was at once seen that
there was a great deficiency of oxygen, and on adding potassic iodide
to the acid which had been electrolysed a considerable quantity of
iodine was liberated. This was at the time supposed to be hydroxyl,
and means were adopted to prevent its being carried by the oxygen
into the potassic iodide used for absorbing the ozone. For this pur-
pose a tube containing potassic dichromate was interpolated, but after
it had been in use for some time it was noticed that the dichromate
was entirely unaffected, and on testing the acid with potassic dichro-
mate and ether it was found that the oxidising substance is not
hydroxyl. Shortly after this observation, Mr. Madan called my
attention to a paper by Brodie, on “ Organic Peroxides theoretically
considered ” (Chem. Soc. J. (1864), 17, 281—294), at the end of
which is the statement that, during the electrolysis of dilute sul-
phuric acid, an oxidising substance is produced which does not give
the reactions of peroxide of hydrogen, but which may be the hydrate
of SO,, or H,SO;. In all probability, the compound is Berthelot’s per-
sulphuric acid, the anhydride of which he obtained by the action of
the electric effleave on a mixture of sulphurous anhydride and
oxygen. I have not yet examined this compound, but I may men-
tion that during a visit to my laboratory, Mr. C. E. Groves found that
after boiling the liquid, it produced perchromic acid when shaken
with potassic dichromate and ether.

In the early experiments, an error was probably introduced in the
determination of the quantity of ozone by the employment of an acid
solution of potassic iodide. Brodie showed that a strong solution of
hydriodic acid gives results much too high, and although my solution
was dilute and, in fact, almost of the strength that Brodie found to
give nearly correct results, yet it will be safer to reject all the
numbers, although it involves the sacrifice of 27 experiments.

The apparatus with which all the recent experiments have been made
is the following :—A U-tube (see Fig., p. 595), 16 mm. in diameter,
has a bulb blown on one limb, aud joined near the top of each limb is a
tube to conduct away the gases. The erds of the U-tube are closed
with corks. The positive electrode consists of a tube closed at the
lower end into which the fine platinum wires are sealed; this tube is
filled with mercury to make the electric contact, and it is necessary
to boil the mercury in the tube to ensure contact with the fine wires.

“aZIS Ivay t ‘dIOY O1HUDHATAG AO SNOILATOG JO SISATOULOATA AHL uO GAsaA SALVUVddYy

596 MCLEUD ON THE ELECTROLYSIS OF

This electrode passes through a slightly wider tube fitted to a cork
well soaked in melted paraffin and fixed to the limb of the U-tube
on which the bulb is blown, the fine wires being about the level of the
bottom of the bulb, The negative electrode consists of a platinum
plate welded to a platinum wire which is also fused into a glass tube
containing mercury and fitted by a cork to the other limb of the
U-tube. The hydrogen passes through the tubulure to a delivery-
tube dipping into a pneumatic trough, and is collected in a graduated
cylinder. During the electrolysis, the U-tube is placed in a beaker
containing a mixture of ice and water which is kept stirred by
blowing air bubbles into it through a thistle funnel as suggested
many years ago by Harcourt and Esson; the air is supplied from‘a
water-pump, and by this device the apparatus may be left with very
little attention. The object of the bulb on the U-tube is to cause the
cold convection currents to impinge on the positive electrode and pre-
vent its rising in temperature. The ozonised oxygen passes through a
narrow U-shaped tube to the absorption-tube containing solution of
potassic iodide, this latter tube being also provided with a narrow
U-tube ; the free limbs of the U-tubes are covered by a closed piece
of glass tube, and the whole stands in a wide test-tube containing
water, thus forming a water joint which has been found more con-
venient than the paraffin joint used by Brodie. The absorption-tube
is made of a piece of wide tube with a bulb at each end; it is placed
in a slightly inclined position, and is similar to some of those
employed in Brodie’s experiments. To the further end of the
absorption tube, a delivery-tube passing to a pneumatic trough is
fixed by a caoutchouc joint, and the oxygen is collected in a 100 c.c.
measure. The troughs are shallow tin vessels filled to the brim and
of such a depth that the pressure is equal on the liquid in both limbs
of the U-tube ; as the water is displaced from the graduated measure
it flows over the edges of the troughs and thus the pressure remains
constant during the whole experiment. The absorption-tube contains
about 30 c.c. of a solution of potassic iodide containing 0°75 gram
of the salt.

When about 100 c.c. of oxygen have been collected, the electric
contact is broken immediately after a bubble has passed through the
absorption-tube, the delivery-tubes are removed from the troughs and
a tube from an aspirator is connected to the oxygen delivery-tube.
The test-tube of the water joint is lowered, the negative electrode
removed and the greater part of the acid in the U-tube is withdrawn
by means of a pipette. This liquid is saturated with ozone and
usually effervesces. A slow stream of air is now drawn through the
apparatus for about half-an-hour, so that all the ozone may be
deposited in the potassic iodide solution. The apparatus is then dis-

100

PERCENTAGE OF OZONE
IN OXYGEN. *>

4
ATOMS OF OXYGEN IN

THE FORM OF OZONE
FOR 100 MOLECULES
OF HYOROGEN.

RESULTS OBTAINED WITH ELI

+——

as 4
/
+ eee
- i
_+-----+----- —

ieee sel }
|RATIO OF GASES
T i T 1

1:025 105.1075 11 1:15 1'2 1:25

DENS

Journ. Chem.Soc. A uqusl_ I886.

WITH ELECTRODES MADE WITH WIRES :045 MM. IN DIAMETER.

TTAL ATOMS |OF OXYGEN FOR 100 MOLECULES OF

HYDROGEN.

|
ATOMS OF DXYGEN IN THE FORM OF ORDINARY GXYGEN FOR f100 MOLECULES OF HYDAQGEN.
|
|
}
|
|
|
|
|
|
———« ATOMS OF OXYGEN IN THE FORM oF “ApTive OXYGEN’ FOR 100 MOLECULES GF HYDROGER.
| a
4 Pi
| / ”
nd

en
a’ ‘
4|~
ae mm
——— sf , a
/ —
— »
\
\
hee 4. \
a etn. ws iemtetieietien weet oot -— Pp
1:25 ng 1°35 4 Le) 5 155 16 1°65 ia

DENSITIES OF SOLUTIONS OF SULPHURIC ACID.

Harrison & Sons. Lith. S' Martins Lane. W.C

RESULTS OBTAINED WITH ELECTRODES MAD

_
100 =
ToT
r
80
ATOMS OF OKYGEN IN THE FORM OF wmen |
4
70
|
—
| i
60 | a
EE Ee
40
= oe
30 4
20 /
PERCENTAGE OF OZONE IN OXYGEN. yj ns
-*K, ae
ATOMS OF OXYGEN IN THE FORM OF OZONE ae 4
FOR 100 MOLECULES OF HYDROGEN or
10 } | —
/ a ite:
\
j ™~
,
o-a@aeoeooor™= + ee r rat Poe
= a2, OF GASE.
0 L__| |
1025 105 1-075 11 1-15 1-2 1

DENSITIES OF SOLUTIO

Journ. Chem.Soe. August /SS6.

DES MADE WITH WOLLASTON WIRES -027 MM. IN DIAMETER.

Pane es
a aa (

TOTAL ATOMS OF OXYGEN FOR joo MOLECULES OF HYORQGEN.

1 : —E

-
}
|
|
|
J
|
PROINARY OXYGEN FOR 1qp MOLECULE OF HYDROGEN.
bn
|
|
|

—+

2
“)

TOMS OF OXYGEN IN THE FORM OF ACTIVE OxyYag m,ron 100 MOLECULES OF HYDROGEN.
t if

|
|

2 1:25 13

5S OF SOLUTIONS OF SULPHURIC ACID.

Harrison & Sons. Lith, S! Martins Lane. W.©

AQUEOUS SOLUTIONS OF SULPHURIC ACID. 597

mounted, the remaining acid poured from the U-tube and measured.
The contents of the absorption-tube are washed into a flask and
mixed with about one-fourth of its volume of dilute hydrochloric
acid, when the brown colour of the liquid is much intensified. It is
next decolorised by decinormal solution of sodic thiosulphate: the
addition of starch is unnecessary, as the disappearance of the brown
colour is quite sharply defined. The gases are measured by placing the
cylinders in water so that the levels are equalised, the temperature
and barometer being simultaneously read. To determine the quantity
of oxygen which has formed the oxidising material produced in the
acid, 10 c.e. of the liquid are placed in a flask, mixed with water and
solution of potassic iodide, and kept in the dark until the following
day, as the oxidising action is ineomplete until after the lapse of some
hours; it is then decolorised by decinormal solution of sodic thio-
sulphate. The oxidising substance not being identified, the quantities
obtained are called “active oxygen,” a name used by Berthelot.

The currents used for the electrolysis are measured by a tangent
galvanometer, the values of the deflections of which are known.
Latterly the wires of the positive electrode have been measured by a
microscope micrometer before and after each experiment; and from
the mean of the measurements, the area of the electrode is calculated,
this combined with the current measured by the galvanometer gives
the current density, that is, the intensity of the current at the surface
of the positive electrode.

The solutions of sulphuric acid were made by diluting with dis-
tilled water some pure redistilled sulphuric acid supplied by Messrs.
Hopkin and Williams. Their densities were taken by hydrometers at
15° C., and the quantities of acid they contained were determined by
a standard solution of sodic hydrate. The following solutions were

used :—

Percentage of f | Molecules of
Density. entag OH, to 1 of | Density. SO.Ho | OH; to 1 of
2109.

:
Molecules of Percentage o

| SO,Hoz.. SO.Ho,. SO0,Ho,.

46 °459
51 333
55 °421
60-978
65 *755
69 *825
74°440
78 *592

|
“92 133°
| ‘373 68°

“JT Ot bo
o or

*122 43 °
4°611 31°
‘061 20
533 14°
°272 10
"105 8

me DS bo COP OI OD

fat mt fed df fd fed feed
wrmnmerEddo6de
Hh Om OOO ES
ee
TIAMAAL god
nr
Bm 0 03 OH Cd
SSRSERB™"*

Various forms of electrode have been tried. In some of the pre-
liminary experiments, pieces of Wollaston wire about 0°03 mm. in

VOL. XLIX. 28

598 MCLEOD ON THE ELECTROLYSIS OF

diameter were used. In others, some drawn wires about 0°045 mm. in
diameter and flattened, were fused into the glass. One electrode was
made with six wires, 0°047 mm. in diameter, the total. length of the
wires being 104mm. A series of experiments was made with this
electrode, the wires being shortened after each electrolysis, until the
total length was only 6 mm. As the wires were shortened the
quantity of ozone produced increased, the current being approximately
the same in all cases, so that it appears that the formation of ozone
depends on current density.

In the first set of experiments of which the results are given below,
an electrode was used consisting of six wires, each about 1 mm. long
and 0°045 mm. in diameter. Duplicate experiments were made, com-
mencing with acid of density 1025 and proceeding to that of 1°35.
In the last experiment, the current was noticed to be very irregular,
which was attributed to a bad contact in the battery, but after
cleaning the binding screws there was no improvement, and the
experiment was stopped. On examining the electrode it was
found that the wires had been much attacked, three of them had
almost disappeared, and the remainder were very much thinned.
The negative electrode was blackened at its lower edge, which was
thought at first to be due to the precipitation of lead from the sul-
phuric acid, but boiling nitric acid did not remove the stain: on
heating the plate red hot, the black colour was at once destroyed and
the metal became lustrous, so that the platinum dissolved from the
positive electrode had been partially deposited on the negative. It
was afterwards noticed that in determining the “active oxygen,” in
the acid that had been electrolysed, after the brown colour of the
iodine had been discharged and before the liquid was quite decolorised,
it assumed the pink colour of platinic iodide.

The next electrode was made with platinum wire containing 15 per
cent. of iridium, this dissolved in the same manner in the acid of
density 1°35.

As the electrodes with fine wires are somewhat troublesome to
make, it was hoped that a thicker wire, but of such a length as to
expose the same surface and thus maintain the same current density,
would be equally efficacious. An electrode was therefore made with
w piece of wire, 0°33 mm. in diameter and 0°6 mm. long, but this gave
less ozone, although the “ active oxygen” was increased.

The next electrode consisted of a fine platinum-iridium wire fed
through a capillary tube, so that a fresh piece could be used for each
experiment. The resistance of the wire, however, made the operation
very slow and the electrode was abandoned.

A new electrode was then made with a single wire 6 mm. long, and
placed vertically in the tube. This gave less ozone; but a similar

AQUEOUS SOLUTIONS OF SULPHURIC ACID. 599

electrode with the wire horizontal produced a larger quantity. This
variation seems to be due to the heat generated in the wire by the
current : when the wire is vertical, the evolved gas travels up the
wire, and if this is heated some of the ozone will be destroyed ; with
the horizontal wire, however; the gas at once passes into the cold
liquid and the ozone is preserved.

The next electrodes consisted of two wires projecting about 3 mm. ;
it appeared occasionally that one of the wires made a better contact
with the mercury than the other, so afterwards a single wire was
passed across the tube and projected about 3 mm. on each side. This
form of electrode was used for the remaining acids, a new one being
made for each pair of experiments.

Another series of experiments with acids of densities from 1°025 to
1:25, was made with an electrode of one Wollaston wire 0°027 mm. in
diameter projecting on each side of the tube, having a total length of
53 mm. This was not much attacked by the acid, and I began to
think that the previous destruction of the electrodes was due to the
presence of a trace of chlorine in the distilled water, but when the
acid of density 1°3 was used, the platinum was dissolved, and with
the 1:35 acid the electrode had almost disappeared, when 53 c.c. of
oxygen had been collected. New electrodes were then made for each
experiment, but the series was not continued beyond the acid of
density 1°5, as the results were so similar to those obtained with the
thicker wires.

As electrolysis seems a convenient process for obtaining ozonised
oxygen, it appeared to be worth while to endeavour to construct an
apparatus which would yield it more rapidly than the one above
described, which has never produced 100 c.c. in less than an hour,
and has usually taken very much longer. Two things are essential,
a high current density and the electrode being kept cool. The latter
condition was fulfilled by blowing a hole on the side of a U-tube, and
fusing on to it a small circular electrode made of platinum foil, and
surrounding the edge of the platinum disc with arsenic glass or white
enamel, which was recommended to me some years ago by Mr.
Crookes. The enamel is very fusible, aud as it adheres strongly to
the platinum and fuses well into the glass, it retains the platinum
firmly in its place. When the U-tube is placed in ice and water, the
heat produced is rapidly conducted away through the plate of metal.
The area of the electrode exposed inside the tube was about 1‘°9 mm.
in diameter, and it was not found possible to obtain a high enough
current density with the U-tube employed, for the resistance of the
acid would not allow a sufficient quantity of electricity to pass. The
result was that the percentage of ozone was only about half that
obtained from the same acid with the fine wire electrodes. No doubt
282

600 MCLEOD ON THE ELECTROLYSIS OF

if a wider U-tube had been employed, the results would have been
more favourable.

It may be well to give an example of the observations and calcula-
tions made in one experiment which will show all the data from
which the results are obtained. I select the one giving the maximum
quantity of ozone when the first electrode was used.

Jan. 12, 1886. Electrolysis of acid of density 1-075.

Electrode of six platinum wires 6 mm. long and 0°045 mm.
diameter.

Area of electrode at sides = 0°045 x 7 x 6 = 0°8482 sq. mm.

“ . ends = 0°045* x z x 6 = 0°00954 sq. mm.

Total area = 0°85774 =
0°0085774 sq. em.
8 cells of Grove used.
Started 9.45, galvanometer deflection 11°
9.50 n ; 11°2°
9.54 ” 11°3
10.0 - 11°4
10.40 116
Stopped 11.7

Mean deflection 11°3°; current = 0°4294 ampére.
Duration of experiment = 82 min.
Aspiration started at 11.12; stopped 11.38.
59°7 c.c. of acid in U-tube.
10°5 c.c. of acid mixed with potassic iodide solution.
Decolorised next morning; 7°0 c.c. of decinormal solution of sodic
thiosulphate required (1 c.c. of solution = 0°000788 gram of oxygen).

0°005516 x 59°7
10°5

7-0 x 0:000788 = 0°005516. = 0°031362 gram

* active oxygen.”
98:1 c.c. of oxygen were collected and measured at 9° C., and bar.
758°3 mm.
261 c.c. of hydrogen were collected and measured at 9°, and bar.
758°3 mm.
7583 ,
8°6 pressure of aqueous vapour at 9°.
749°7
98:1 x 749°7
760 x (1 + [0°003665 x 9])
760 mm.

= 93°68 c.c. of oxygen at 0°, and

AQUEOUS SOLUTIONS OF SULPHURIC ACID. 601

261 x 749°7
760 x (1 + [0°003665 x 9])
760 mm.
0°09368 x 1:4336 = 0°13430 gram of oxygen.
02492 x 0°0896 = 0:022328 gram of hydrogen.

The liquid from the ozone absorption-tube required 10°05 c.c. of
decinormal solution of sodic thiosulphate to decolorise it (1 c.c. =
0002366 gram ozone).

10°05 x 0°002366: = 0°02378 gram ozone.

Of the ozone passing into the potassic iodide solution, 4 of the
oxygen is fixed, and % passes on and is collected with the ordinary
oxygen.

= 249°2 c.c. of hydrogen at 0°, and

0-02378 ; = 0-01585.

Total oxygen = 013430 gram.
0°01585

Ordinary oxygen = 0°11845 gram.

This ordinary oxygen was mixed with the ozone as it left the electro-
lysis apparatus.
0°11845
0°02378
0°14223
0°02378 « 200 _ 42. :
“aa 16°719 per cent. ozone by weight.

In order to make the results of the different experiments com-
parable, the various products are reduced to molecules, and compared
with 100 molecules of hydrogen which is the only constant.

0:022328
2
0°11845
16

= 0°011164 molecule of hydrogen.
= 0°007402 atoms of oxygen in the form of ordinary oxygen.

= 0°001485 - ozone.

0-031262
16
0:011164
0011164

= 0001960 ~ ** active oxygen.”

x 100 = 100 molecules of hydrogen.

0007402 x 100 = 66°302 atoms of oxygen in the form of ordinary

602 MCLEOD ON THE ELECTROLYSIS OF

0-001485
0011164
0°001960
0-U11164

x 100 = 13°356 atoms of oxygen in the form of ozone.

x 100 = 17°556 a : “ active
oxygen.”

The total number of atoms of oxygen for 100 molecules of hydrogen
should obviously be 100.
66°302
13°356
17°556

97°214

The total is usually a little less than 100 (occasionally a little above),
this loss seems due to absorption of oxygen by the water in which it
is collected.

The ratio of the gases is obtained by dividing the volume of hydro-
gen collected by the quantity of oxygen. The volume of the oxygen
collected being equal to the volume of ordinary oxygen plus the volume
of ozone—if our views of the constitution of ozone are correct.

249°2 _

= 2°660.
93°68 "

Ratio of gases =
249°2

‘
—

Quantity of hydrogen evolved per minute = = 3°039 c.c.

1 ampére evolves 6°9549 c.c. of hydrogen per minute. The current
used in the experiment was 0°4294 ampére—

0°4294 x 69549 = 2°986 c.c. of hydrogen per minute.

This number is a little less than that observed, doubtless from the
mean deflection taken not being the exact one.
If we had taken 11°5° as the deflection of the galvanometer, the
current would have been 0°4372 ampére, and—
0°4372 x 69549 = 3:034 c.c. of hydrogen per minute,
which is almost identical with the measured quantity, but the
previous number is sufficiently accurate.
The current density is calculated by dividing the current by the
area of the electrode in square centimetres :—
_ 0°4294
0°0085774
This example will render intelligible the following table of collected
results. It has been considered advisable to give the method and
numbers in detail, for at present it seems difficult to see what bearing
they have on the phenomenon of electrolysis.

= 50°06 ampéres per square centimetre.

‘MOT 00} FONUI og 4SNUI sioquINU 1044%] 043 ‘syUOMTIOdx9 07} [[@ SutsMp AoJOUIVIP WAOFIUN B pel] SALA
04} 984} UoI4duInsse 944 UO pozeNOTVd oA SaIJISUOP JUAIIND 94} SY ‘“poajossip useq savy 07 PUNO; sv OpO.r}O9/T

é 5 ZL48-2| 16L-L |Z1S-O0T FHG6- 81) 88Z-9 | O&8-6L| OOT
L-8h | ¢ -0| 16-2 | SZh-Z | S96- OT |LEI- 101, 98P- ST | S68-6 | 8SZ-9Z | OOT
8-2 b-0| $22.8 | Z6F-Z | LG8- FL |G88-86 | OSF- FL | GES. ZT | 988- TZ | OOT
SI. #¢ -O| 822-8 | 80S-Z | 809. FT |EZT- 86 | OZE- FL | EFZ- ZT | Z9S-TL | OO1
Cl. 6F 1Z6-Z | 660-€|#6S-6 [814-26 | 986-62|86F-9 | $8z-19| OOT
08. LF 916-2 | 180-€| STF-6 |99F-86 | 98F-1E/88z-9 | 269-09
90. 0¢ , 966-% | €81-8| 416-6 |66Z-66 | ZZ. 8E | ShS.9 | P8E-6¢S
9Z-8¢ 008-€ | 2460-€| 1ZF-6 |1Z0- 101] F88-*E|62z-9 | 80F-09
8T- FS ¥2Z-@ | OFO-£ | Z90-€T [299-96 8614.12 |Z86-8 | Z8L-6S
€9-*S €hI-S | P80. €| 910- ST \SEZ-F6 | 8S8-Sz|So6-8 | 026-6¢
z-99 , 900. % | 824-2 | 669- #1 |\49L- 66 | 9SS. OZ | BEI. 69¢- 19
9Z-&¢ , €91-€ | 8ZL-Z | EL8- ST |S88-96 | F6L- 0Z | G00. 889. $9
90- 0S P6ZF-0| 680-¢ | 099- 614-91 |F1Z- Z6 | 9SS- LL | 9S¢. Z08- 99
£9. FF -% | 8288-0] 09.2 | 6S9. Z| 899- 9T |849-26 | 940-81 | 992. 9&8. 99
€6-9¢ -€ | $88PF-0]} 192.8 806- FT \699-96 | ZL0- TT | ZPL-21 | 0S8-2L
26-0¢ -§ | ZL8F-0| S80.¢ | ze P8T-ST |\LE6-96 | 0Z0- IT | ego. P98- ZL
86- 2 628Z-0| 896-1 | IPE. Z| 68E-ET \SS¢.96 | 8SI-9 | S¢6. CPP. LL
94-08 689Z-0| €&8-T I9F- ST 828-246 | O8T-9 | ChE. 8h8-6L

10
”

Ce ee cee i Be oe eo ee Oe Oe Oe oe ee

ite)

19 19

‘99 ** IQA9UIRIP “UOT CFO.
puv Suc ‘wut, ynoqvsaim =
wnuyed xIs Jo sporqoopsy

1d uw

ST
€1
rat
Il
€1
ZI
&1
ZI ‘uve

SSS SSS RH AANAN OO

‘wo ‘bs |

10 “9° ‘uas {xo |

d “poyy a0 uss 4x0 |

sored ‘miut | ‘sased ul =| ‘sur10798
. -n9o[Bo . | SE bee ‘prow jo

-wy ‘ aad | yo | ouozo | uad{xo 7 se

; utur [Pomere _— | A4q1isueq

quar pr PeAaToaod O1gey | Jo-quaa | jo uN,

“1nd Jo H log | jo w40F

Ayisueg H aq} ur uasAxO Jo sMIOTY

(=)
_
3)
=
2)
_—
ce
=)
a8 |
Ay
H
=)
77)
Fy
oO
N
Zz
=)
Lani
=
=)
|
io)
mM
RQ
~
)
i)
>
g
=

eAlpY ,,

Arvutpio

i
©
mM
La)
RD
bn
_
©
=]
e
oS
:
=
=
e
Z
jo)
o
iS)
~
=
oO
=

‘AIIM JO PUD TSA »

LY -&F
FL-09

6% -8ES\d

cc. 6S
9F- LS
T9- 6P
@S-0¢

L8Z-Z
98Z- €
TtO- €
ZEG- Z
609- Z
OLL-%
689-2

LIL-%
SLP-&

68Z8- 0
CZL¥P- 0}
ZLEF- 0}
91Zb-0
1¢Le-0
€868- 0}
1988-0
9068-0
100: 0

6922-0
I1¥Z-0

0Z- 3%
18Z-
Z10- &
6L0-€
909-@
LVL.Z
08-3

GES: Z
€29-€

L09.-T
¢L19-T

6¥0-%
£00.

€Z0-&
OIL &
Ch0. 8
820. ¢ |
eF0- 8

NOnaa On

L2g.
L¥8 -
Lz.
ECP:
PG6E-
v6P- IL

LLZ. OT
840-6

F06- L
90-9

619-6
8€9- OT
bF6-PL
189. 6
684. &T

216: 86
LOE. 66
O1Z- 66
080. Z6
€EP- 66

ChO. 86
$63 - 86

906. 96
698-46

692-86
£0Z -86
698 -L6
OF0- 66
699. L6

889- 66 |

112-86 |
8ZP- 66 |

|
90¢- LT [€T-OT
98b- SL BhE-6
| |
848- 9% | 940.

Z28- 16 | 69-3

|
8¢0. 91 | 128: £
860: 91 | LFZ-8
Z49-F1 | CFS-Z1
SLT. 88 | 228-9
29-61 | 644-01

266. FL
PES- EL
o89. ZL
O6F- OL
940. TL
98¢-TZ

681-69
920- 69

08S: OL
OS9- TZ

2°26 - $9
8LE- €9

068. FZ
8c8. €Z
2ve- OL
O6F- 6¢
898. 19

ee ee eecaces 9po1yz0e]9 aeyrung {

YOR ‘SOITM OMY TTA OPOLZOOTH
cineca 488] oll 9po140e[9 MON
te eeeees se reUOZLIOW OTA
qnq ‘488 exIT S9porqoeTo nt
Feet ee ee eeeeeeseees THOT
-I9A “AOJOUIVIP “UU CHO.
‘Suol ‘uit g odIM wmnIp
“Wat wnuryed eposqoota MON
tet eeeeeeeeeseeeses gang
Aaeqypideo = ySnoryy = pay
‘Suo] “wut g o41M UINIp
“at urnuiyeyd jo eposoe[iy
rhteeeeeeseeees JQOUIUIp
‘mm ¢¢.9 { Zuo, ‘mu 9.9
OMIM YOIYZ JLOYS JO OPOIpoTT

teeeeeeeeeeeees gonouruIp
‘MUL CFO.O pus Fuoy ‘uu T
qNogesetim aUy XIg “epory
-o9]9 UINIpIt umnuyeyd soy

“mo “bs
aod
sored
“ay
“quad
-1nd JO
fysusog

-wy
“quod

-Ing

o°9
‘ulm | ‘sosed
aod jo
pesjoao| o1gey
H

‘ues {xo

UY
9u0zo
jo"jue9
1og

*gUL078
| wad 4x0
| jo ung

‘uo8fxo
£1Buipig

jo uLtoy
ey} ut ues {xo jo

suloyy

“uss01pAq
jo sonoe[0W

“pros jo
Ayisueg

“poyonyye Tonur epolzqo[T #

eporyoo]o ouIeg

eporyoojo ouleg

eporyo9]o oULeg

epoiqoaya ouleg

* epoiqoo]o oureg

toteeeeeeees gposqoota ouneg

tree eererees Qnoiqoa]o OUIEG

tHeeeeeeeses gnosao[e ouleg
‘UU £70.09 “WuBIp “WUT ¢.g |

qnoqe yysue, B40} ‘sorLM

UOWSVTOAA OMJ JO OporzOOT

te oeeeeeeees gpoaqoate ouleg

so F898" Qpougoale LBTIUIG

*** gpoaqooje aepIUIIg {

#66: ZOT T6L6-0| 826-1 | Z8b- Z| PLL- ST |FS9-86 | T8T-FL | O19- IL 298-22
GZ. 64 689Z-0| S64-T | LL1-€| 99-8 |Z1Z- LOT) 96F-9E|OSS-G 9Z66-6¢
cS. 92 LL9Z-0)\ FOE-T | Z61-& | Sh6- OT |STF- 26 | USP-ZE| 196-9 O06. LS
gc .94 LL9Z-0| Sh8-T | 886- Z| 9EZ-F1 \6EF- L6 | 961-22 | 666-6 $92-09
8¢- ¢8 CEIS-0)| 10Z-% | 6IL-Z | L2S- 91 |00Z- FO1| Z8L- L241 | 9P8-ZT ZLS-EL
Z8- $9 I1¥Z-0| EL9-T | PEL. Z| Z8E- LT |96S- L6 | 984-06 | 8EE- ET | ZL¥- €9
18. 64 6S08-0| F8I-S | G8-@ | ELS. OT |20€-S6 | SIS- HL | 68P- ST | 90E. 69
69-99 G1L4Z-0| €98-T | 68b- Z| 660. LT |ELH- 96 | 9FZ- TL | €S9- FL | 19-02

Vx)

Se ANNO

ite)
2

See eee ee

oS. 6P ChIZ-0| ZLF-T | S98-Z| POF- ST |S8T-S6 | 990-9 €14-€1 | 90F-SZ
36-FL 8860-0] 489-0 | 8Z8.2| 828-0 |6L6-86 | Z19-F1 | 669-0 | 899-8
9€-12 G0ST-0| SO. T | Lob-2| 99-0 |FZ8-66 | 02-81 |8zS-0 | 920-8
ZL- 9% 9YT9T-0| SE1-T | 81e-2|9F9-0 |STT-86 | 8&E-ST | 919-0 ~9Z-6L
LG- SG LOLT-0| 98Z- T |S1S-2| 189-0 |6248- 86 | Z61-61 | ShS-0 | PHT-62
86-1F LLG. LL8-T | L29-2| F8L-0 |8S¢-66 | $9Z-2'ec¢-.0 | 98L-SL
LS. ¥E LL9Z-0| 898-T |8ES-Z| 128-0 |Zoh-66 | O&F- 02 | 8t9-0 | FLE-8L
60- GP 0208-0| 68F-Z | 909.2| 966-0 |L0Z-66 | 816-62 | G9L-0 | 22-92
€F- VE 692. 6LL-T | 009-2 | 086. 0 |S&P-TOT 08Z-%2| 212-0 | 8Eh-9L
99. LP 18h PEE-% | O&9-2| €9b-T |P8S-OOT| Z8L- FS | 9TI-T | 982-92
18: 6S . L426-% | G8Z.2|919-T |90P- TOL 812-62 | F9T- 1 | #20- 12

Ne)
wMeooMm
e

0 16

SF 8

*** gporz0a]o aoyrag {

**** gporqoaye rBpIMTg {

Re ee oe ee

A asm aig cana, Ad Ne

*** ** aporqoojo sexys {
‘opis qoR
uo ‘wu g ynoge Suyool
-O1d OtIM OUO ALM Oposoa[y

uw 1

99. SZ 6&2S-0| 8Sh-€ | Sze.Z| 8S8-8 |1Zh- 66 | 808 - ZT | $€z- 4 08s -

| &
?

‘wo ‘bs oo
aod : vio ‘o"0 ‘uaa {xo
sored peau! “UT ul “71048 |
-noyeo . : 7
-uly mae aod euozo | ueshxo
“quad aa peaoao jo "quad | Jo ung |
-md Jo H 19g jo UlI0F
fyisueg H (yg ut uesk{xo jo sulo}yy
|

“UadOIpAy
JO sapnos[OW

)
<

i.)

—

rm
=
=
Ay
<
5
mM
&
3
2)
Z
S
lam)
=
=
<
Ss)
NM
2]
=
Ss)
&
‘=
o
<

freutpio

“pealosstp epo1zo9[

. 61010.0 ‘BAY "88Z.9° yaSu9]
91.9% | FIS-€ |SOLF-0| FEE-8 SFL-Z| OTL. FI |L48- 96 | OFE- Oz | E82. TT | Fz. 29 a: ae oe men
| *SOUIM J.LOYS OT WITM Oporpoory
000008 06505008: ge oe
29.4% | 998-F |Z869-0| 866-6 OTS-%| 689-4 |¢24-26 |8e0-9T| 192-9 | 94F- SL T- eps wo pom apn oun
SI-ST | 986-6 |¥oz9-0] ¥65-6 | Fee. Z| S264 |989-86 |OL8- 01 | Zz8-9 | 9FF- 18 ‘ae G1 MOGU apON}DATD Ly
| ** TUT . “urerp § ZZ
L-G€ | OL1-T | I69T- 1-0) TIé.T | 88- n| ene. a 2 POL FI6- 61 | S06-T | 099-28 at 810.0 colnen ae
bas P UU 910.9 “uUeT ce
€-L81| TL8-T | 6822-0) 909-1 | F09-<| te ® |00L-66 299-81 | 918-8 | ZEB. LL +14 | ugSuay fl “opasioaye aon

** "UU OZ0.4) “ULB A
LV. @ | 680-8 [286 OO WE. LT | 869-9 | 816-94 . ‘pduay [WY paLoye AON

¢.9F1) 120-2 | 9062-0) L6L

#&6-€S1| $10.2 | 2e6z- 0) 920

** “UUUL OZ “TUR A
829-2 | 6LE- ZI ps 86 ba ST | 008-01 S98. &L T 4 lates wane uear a: Ay mat re Aon

m
co
hh
_—
n
-
W
io)
[om
e
)
=
=)
a
2
is)
e
7
©
i=
oO
te
—
Ss)
=

“m0 “bs
i) |
aod 0"
‘pois, | ‘sored
sored inn te ‘mut =| ‘sased ur | “su07"
“wy : ot . ‘V aod jo | euo0zo | uat{xo
“quad we — JATOAO | O198 0 “4ua0 | Jo ung
yu: sod | -ang |Ps1042| ovyey | Jo"u20| Jo ung
-In9 Jo H H 19g | }O WLI0F
syisueg oy} ut ued {xo jo sul0yy

‘aod fxo

|
|

AavutpiQ

‘uas {xo

“prior jo

Ay yisueqd

‘uasorpAy
jo Se

AQUEOUS SOLUTIONS OF SULPHURIC ACID. 607

Various estimates of the solubility of ozone in water have been
made, and one experiment has been carried out to endeavour to
determine the coefficient. The result is so very different from some
of those previously obtained, that it must be regarded as only pre-
liminary, and other experiments must be made to confirm it.

The tube used for the absorption of the ozone in the above experi-
ments was filled with water and connected to the electrolysis apparatus
containing acid of density 1°075, and a current of electricity was then
passed through the acid for 232 minutes, during which time 500 c.c.
of hydrogen were evolved. From other experiments, the oxygen
may be supposed to have contained about 16 per cent. by weight of
ozone, or 11°27 per cent. by volume.

The water in the absorption-tube, which was at the temperature of
14° during the experiment, was mixed with potassic iodide and hydro-
chloric acid, and decolorised by the decinormal sodic thiosulphate.
0°9 c.c. were required equal to 0°002129 gram of ozone. The volume
of the water in the tube was 32 c.c., so that 1 ¢.c. of water contained
0°002129 _ 9.00006653 f

—— a 53 gram of ozone.
1 litre of ozone weighs 2°1504 grams, and
000006653 x 1000
2° 1504

= 0°03093 c.c. of ozone in 1 c.c. of water.

The pressure of the ozone in the tube, calculated from the volume
100

of an atmosphere for
11°27

per cent. of ozone, was = 8875, so

QO”

75
that the coefficient of solubility of ozone from this single observation
is 0°03093 x 8875 = 0°2745, which is nearly ten times that of
ordinary oxygen, 0°02989.

The coefficient obtained by Schéne (Ber., 6, 1224—1229) was 0°366
at 182°. That obtained by Carius (Annalen, 174, 30) was 0°834
at 1°,

When ozone is passed into a solution of potassic iodide, especially
an acid solution, a dense white fume is produced. This is generally
said to be hydric peroxide, but I had a suspicion that it at least con-
tained iodic acid. An attempt to prove this was, however, a failure,
owing to the omission of a precaution which was obvious when it was
too late; however, since the paper was communicated to the Society
I have had an opportunity of repeating the experiments in a more
careful manner.

About 100 c.c. of ozonised oxygen, containing probably about 15 per
cent. of ozone, were collected in a small flask. When the flask was
filled, a solution of potassic iodide mixed with about one-third of its
volume of dilute hydrochloric acid was introduced into the flask and

608 ELECTROLYSIS OF AQUEOUS SOLUTIONS OF SULPHURIC ACID.

agitated with the gas. After a minute or two some sodic hydrate
solution was introduced into the flask to fix the free iodine (this was
the precaution omitted in the previous experiment), and after agitation
the flask was fitted with a cork carrying two tubes, and one of these
was attached to a glass tube heated in the centre to the softening
point. Between the flask and the hot part of the tube were placed a
slip of paper coated with starch, and next a piece of paper with
potassic iodide and starch. Beyond the hot portion were two similar
papers. Air was slowly drawn through the flask into the tube by an
aspirator. The first piece of starch-paper was entirely unaffected, the
first piece of potassic iodide and starch-paper was rendered very
slightly blue, from the presence of either a trace of ozone or of
peroxide of hydrogen. The starch-paper beyond the hot part of the
tube became intensely blue, from the iodine liberated by the decom-
position of the iodic acid, subsequently the second piece of potassic
iodide and starch-paper became equally blue from the action of the free
iodine on the starch, for of course the high temperature of the glass
would have destroyed ozone or hydroxy] if either had been present.

Note on the Action of Oxygen on Hydroeliloric Acid under the Influ-
ence of Light.—In the foregoing experiments, it was noticed that
when concentrated hydrochloric acid was added to potassic iodide, the
liberation of a small quantity of iodine took place, whilst no such
result followed the addition of dilute acid.. The concentrated acid had
been standing in the laboratory for some time in a bottle only partly
filled.

It struck me that the reaction might be due to the evolution of a
small quantity of chlorine from the: oxidation of the hydrochloric acid,
and it recalled an observation which I had made several times when
preparing a lecture experiment to show the decomposition of water
by chlorine. This experiment is done annually, usually in the winter
months, by filling a flask with chlorine-water, inverting it in chlorine-
water, and placing it in the field for several days; sometimes the
solution is entirely decolorised, and at other times the colour remains
light yellow even after long exposure. During last winter, two
experiments were made, one with some chlorine solution that had
been in the laboratory for some time, and the other with some freshly
prepared chlorine-water. The first was entirely decolorised and the
second was not. This seemed to show that when a certain quantity of
hydrochloric acid has been formed, the decomposition of the water is
arrested, or at least checked; and it seemed probable also that the
process would be reversible, and that hydrochloric acid would be
decomposed by oxygen.

To put this to the test six V-shaped tubes, closed at one end, were
filled with the following liquids ;—

GLADSTONE ON ESSENTIAL OILS. 609

1. Saturated chlorine-water.

2. Chlorine-water mixed with an equal volume of distilled water.

3. Concentrated hydrochloric acid, density 1:153, saturated with
chlorine.

4. Hydrochloric acid, density 1:1, saturated with chlorine.

5. Hydrochloric acid, density 1-03, saturated with chlorine.

6. Concentrated hydrochloric acid with oxygen in the closed limb.

These tubes were exposed to sunlight on May 5, and left in the field
until the 18th. The sunshine was very brilliant on the Sth, 6th, and
parts of the 7th, 8th, 9th, 10th, 15th, and 18th.

The dilute chlorine-water was soon decolorised, oxygen being
evolved; the concentrated solution retained a colour for some days
longer. No gas whatever was evolved from the hydrochloric aci@
solutions that were saturated with chlorine ; and almost the whole of
the oxygen was absorbed by the concentrated acid.

No free chlorine remained in what had been the chlorine-water, as
it gave no coloration with potassic iodide.

When the hydrochloric acid which had absorbed the oxygen was
tested in the same way, a brown coloration was produced.

I am not aware that this fact has been published, at any rate it is
not generally known.*

LVII.—On Essential Oils. Part Ill. Their Specific Refractive and
Dispersive Energy.

By Dr. J. H. Giapsrone, F.R.S.

My former papers on the essential oils (Chem. Soc. J., 17, 1, and
25,1) date back to 1863 and 1872. Since that time these sub-
stances have been made the subject of many investigations by
Riban, Beilstein, Alder-Wright, Armstrong, Tilden, Atkinson and
Yoshida, Pattison-Muir, Flavitzky, and others; many new facts have
been added to our knowledge, and some advance has been made
towards determining the rational constitution of these compounds.

* On the day this paper was read (June 17), there appeared in “ Nature” the
abstract of a communication by M. Leo Backelandt to the Bull, Acad, Roy.
Belgique [3], 11, 194, on the oxidation of hydrochloric acid under the influence of
light, in which it is shown that when concentrated hydrochloric acid is exposed to
sunlight in a badly stoppered bottle, it becomes yellow from liberation of chlorine,

610 GLADSTONE ON ESSENTIAL OILS.

Nevertheless, there remain some most important differences of
opinion on the subject, and eminent chemists are still discussing
whether the formula of oil of turpentine itself should contain one,
two, or three pairs of doubly-linked carbon-atoms. In this paper
I propose to bring forward such arguments as may fairly be
founded on the phenomena of refraction and dispersion, not con-
sidering that this is necessarily the key to the position, but feeling
sure that in the conflict of opinions such arguments ought to have
considerable weight.

Pattison-Muir (Trans., 1880, 688) perceived the value of the argu-
ment from the refraction equivalents, but the essential oils to which
he applied it were of uncertain purity. More recently Kanonnikoff
(J. Russ. Chem. Soc., 15, 434) has attempted to solve by means of it
the constitution of these and many other organic compounds.

Unfortunately very few of those who have worked on these oils
have recorded their refractive index, and still fewer their dispersive
power; and although the specific gravity is generally given, it has
not always been taken at the same temperature as the optical pro-
perties, so that the observations are of little value for the present
purpose. Thus I have to fall back mainly on my own observations,
which are scattered through several papers in the Transactions
of this Society (1870, 101, 147; and 1884, 241), and on some fresh
observations the details of which are given in the Appendix.

It may be desirable in the first place to point out briefly the kind
of evidence which these optical properties afford. The specific
refractive energy of a substance is its refractive index minus unity
divided by its density. It is calculated by me for the line A of the
solar spectrum, and is therefore represented by the formula “4 2 .
The specific dispersive energy is the difference between the specific
refractive energies of the two extreme visible lines A and H, ‘¢.e.,

ny OF a MA, The specific refractive energy mul-

4 gives the refraction equiva-

tiplied by the atomic weight, P

lent. Now the refraction equivalent of a compound body is the sum
of the refraction equivalents of its components; and in all the
saturated carbon-compounds the refraction equivalent of one atom of
carbon is found to be 5:0 for the line A, and that of hydrogen 1°3. It
was, however, observed long ago in the Journal of the Chemical
Society (1870, p. 151), that “all the essential oils of the CyHi. type
that have been examined, more than 30 in number, give refraction
equivalents a little above theory; those of the great turpentine-
group generally exceed the calculated number (70°8) by amounts

GLADSTONE ON ESSENTIAL OILS. 611

varying from two to three, those of the orange-group by amounts
varying from three to four.” It was also observed that the oxygen
compounds of this group afforded “another illustration of the
increased influence on light of carbon which is uncombined with two
atoms of hydrogen or one of oxygen.” Subsequently Briihl showed
that in every compound in which any carbon-atoms are supposed to
be “doubly-linked,” each such carbon-atom has a refraction equiva-
lent of 6:1, or, which is the same thing, each pair increases the normal
equivalent by about 2°2.

It was also an early observation that the increase of refractive
power in the essential oils, aromatic compounds, &c., is attended by a
corresponding advance in dispersive power. This advance is pro-
portionally much larger than in the case of the refraction. Its
extent may likewise be determined by multiplying the specific

dispersive energy by the atomic weight pon". These “ dis-
¢

persion equivalents” have been little investigated as yet, but the law
appears to hold good among analogous compounds that the dispersion
equivalent of a compound is the sum of the dispersion equivalents of
its constituents.* Among saturated organic compounds I find the
dispersion equivalent of CH, to be 0°34 (or more exactly, perhaps,
0°342). It is more difficult to assign an exact value to the two
elements themselves, but 0°045 for each hydrogen, and 0°25 for
carbon may be accepted as a first approximation. Again, just as in
the case of refraction, the double-linking of a pair of carbon-atoms
increases the dispersion equivalent. This increase is greater in the
aromatic than in the allyl compounds, amylene, &c. ; in the latter the
increase is about 0°5 for each pair of carbon-atoms ; in the former it

is fully 0°8.
The Hydrocarbons.

It is well known that the natural essential oils are generally a
mixture of isomeric hydrocarbons of the molecular formula C,H, or
C,sHy, with C,, compounds containing one atom or more of oxygen.
These have to be separated by fractional distillation, and it is difficult,
if not impossible, to obtain the various substances in a state of purity.
It is now generally accepted that the isomeric oils of the formula
C,H. fall into two groups—the terpenes and the citrenes, or isoter-
penes. These two groups, together with the cedrenes, C,;H., differ

* There are two or three considerations which show that this law cannot be as
trustworthy in the case of dispersion equivalents as it seems to be in the case of
refraction equivalents, but it is exact enough for the purpose of this paper. Some
dispersion equivalents were calculated by me 20 years ago: vide Brit. Assoc.
Report, 1866, Trans. of Sections, p. 37.

612 GLADSTONE ON ESSENTIAL OILS.

in boiling point, specific gravity, and rotatory power, and also in
specific refractive and dispersive energy. The difference is strongly
marked in the table given in the first part of this communi-
cation (Chem. Sec. J., 1863, 1), and the average result of these old
observations is given below. The terpenes are taken from nine
different specimens, imcluding turpentine, mint, anise, thyme, and
wormwood: the citrenes from twelve, ineluding orange, lemon,
bergamot, carraway, dill, cascarilla, and elder. The specific dispersive
energy has been added :—

Terpenes. Citrenes.

Boiling point .. 160° C. About 174° C.

Specific gravity at 20°C.......| 0°8617 0 '8466
Refractive index for line A | 1°4612 1 °4636
Dispersion ...-... occce 0 0258 0°0285
Specific refractive energy . -| 0°5361 0°5476
Specific dispersive energy. . we eeee 0 -0299 0 0337

The oils of nutmeg and carraway contained both terpenes and
citrenes mixed together, and of course it was scarcely possible to
separate them. The same appears to be the ease with the oil of sage,
according to Pattison-Muir.

The Terpenes.

From the figures in the above table, it is easy to calculate the
average refraction and dispersion equivalents of the pyre in my
old observations. poe for CyHy, will be 72°91, and ps Ps A will
be 4:07. If, however, we select the three specimens of i ieaile
turpentine, which gives perhaps the purest terpene, we obtain 72°95
and 4°01 respectively. A carefully prepared specimen from American
pine was recently examined, as well as a terpene from oil of mint, the
particulars of both of which will be found in the Appendix. Again,
Flavitsky has determined the optical properties of a left-handed ter-
pene from Russian turpentine, and, though he has measured the lines
a, B, and y of hydrogen, it is easy to calculate his results for the
solar A, and the difference between H and A.*

* This is given in Kanonnikoff’s recent paper (J. pr. Chem., 1885, 497). The
Continental observers generally measure the three lines of the hydrogen spectrum
H., He, H,; of which a and § are coincident with the solar lines C and F. The
dispersion in these observations is ~ — a, which is easily converted into the disper-
sion H — A according to the ratio of 1 to 1°59. The value for A is obtained from
C and F upon the basis of F - C: C — A:: 100: 29°8.

—) — ae bet

GLADSTONE ON ESSENTIAL OILS. 613

The following are the results :—

Refraction Dispersion

‘
Natural terpenes. equivalent. equivalent.
Old observations ............. 72°91 4°07
Three from pine .........+. oe 72°95 4°01
From American pine.......... 73°11 3°99
From Russian pine .........+. 73°12 3°95
From mint... ....seeeseceeees 72°91 4°02

But substances of the composition C\H\, have been prepared arti-
ficially by the action of certain reagents on the natural terpenes
or their compounds. Riban (Compt. rend., 78, 788, and 79, 314)
has determined the refractive indices of his “ terebene” and “ tere-
benthene” for the lines C, b, and others, at 16°C., with the following

results :—

Tere
Sesebene. benthene.
Boiling point .. oc cecces 156° 156°5°
Specific gravity at 20°. coeees 0°8600 0° 8601
Refractive index (C) iiiiccss 1°4645 1°4665
Specific refractive energy. . 0°5383 05406
Specific dispersive energy (b - - C). 0 -0083 0 °0087

The equivalents calculated from these figures for A and H—A are—

Ref. eq Disp. eg
‘ Terebene” .....0..06 72°8 3°81
“ Terebenthene”’...... 73°0 4°00

Barbier gives the refractive index of the CyH,. regenerated from
his liquid CyH,,Cl, and that from another chloride, for the line D.
Unfortunately he does not mention the temperature, but I assume it
to be 12°, which seems probable from his paper. The results are as

follows :—

| Specimen I. | Specimen II.

0°539

Boiling point .. occcces 157° 157°
Specific gravity 8" 08812 0°8724
Refractive index, line D....... 1°4692 | 1 °4704

Specific refractive energy......} 0°532

VOL. XLIX. 21

GLADSTONE ON ESSENTIAL OILS,

These somewhat doubtful data give respectively 72°4 and 73:3
for the line D.

These four artificial hydrocarbons are evidently identical in their
refractive energy as well as in their other physical properties with the
natural terpene, though they have lost the power of rotating the
plane of polarised light.

Now the refraction equivalent of CH, taking all the carbon at
the normal value in saturated bodies, would be 70°8, and the disper-
sion equivalent would be about 3°24. If we add to these figures 2°2
and 0°8 respectively, we obtain numbers almost identical with those
of experiment—

Found. Calculated.
Refraction equivalent ........ 73°0 73:0
Dispersion equivalent........ 40 4°04

being a twofold indication that in the terpenes there is only one pair
of carbon-atoms doubly-linked.

This result, improbable as it may at first appear, only corroborates
the views of those chemists who, on chemical grounds, have held
that a terpene requires the addition of only two atoms of hydrogen,
or their equivalent, to saturate it.

Very interesting also is the solid inactive camphene, CywHy, of
which Professor Tilden has kindly furnished me with two different
preparations. Two solutions of each of these in alcohol were deter-
mined, giving the following results :—

Camphene. Percentage in Refraction Dispersion
solution. equivalent. | equivalent.
lst preparation ..... 28 ‘78 71°82 3°78
9 enone 33 *60 72°17 3°84
2nd preparation..... 29 -69 71°87 3°57
._ - ewe 25°88 71°78 3°47

These four refraction equivalents are as accordant as determina-
tions made from solutions can be expected to be, but they are all
lower than the equivalents obtained from the liquid terpenes. The
same is true of the dispersion equivalents, especially those of the
second preparation. Whatever may be the cause of this, we may
at least safely conclude that camphene does not contain more than
one pair of doubly-linked carbon-atoms ; this agrees with the chemical
character of the compound, as pointed out by Tilden, for it com-
bines with only one molecule of hydrochloric acid, and is directly
converted by oxidation into camphor, CjH,O, which is certainly a
saturated compound.

GLADSTONE ON ESSENTIAL OILS. 615

The Citrenes.

From the table on page 612, it will be seen that the specific re-
fractive energy of the citrenes examined 23 years ago was 0°5476,
and the specific dispersive energy 0°0337. But it must be observed
that some of the hydrocarbons included in the calculation were con-
fessedly difficult to rectify. If the four specimens derived from
orange and lemon-peel be alone taken (which are perhaps the purest
specimens), the specific refractive energy would be 0°5494, and the
specific dispersive energy 0°0331; while a carefully prepared speci-
men of hesperidine from Dr. Wright, more recently examined, gave a
specific refractive energy of 0°5513, and a specific dispersive energy
of 0°0327. Flavitsky also has determined the optical properties of
the hydrocarbon from oil of lemons, giving the specific refractive
energy of 0°5579 for Ha, and the specific dispersive energy of 0°0215
for the hydrogen lines a and y. These figures, reduced as before,

give the following values :—

Natural citrenes. Retre a Dispersion
equivalent. equivalent.
Old observations ..........+.. 74°47 4°58
Four from orange and lemon... 74°72 4°50
Wright's... .cccccccccccccsees 74°98 4°45
Flaviteky’s .. cc cccececcecsecs 75°34 4°55

A citrene also, equally with a terpene, is capable of reproduction
from its compounds. Riban prepared a substance which he termed iso-
terebenthene, having the boiling point 175°, a sp. gr. of 0°838 at 25°,
a refractive index of 1°4677, and a much higher dispersion than his
terebenthene. This gives a specific refractive energy of 0°5531 for
the line C, and a specific dispersive energy of 0°0099 between C and b.

These figures reduced to the probable values for A, and H—A will
give—

Ref. eq. Disp. eq.
Artificial citrene ....... . vol 4°58
Now the increase in both these optical equivalents requires us to

suppose that not one, but two pairs of the carbon-atoms in the
citrenes are doubly-linked. The figures are as follows :-—

Citrenes. Found. Calculated.
Refraction equivalent ........+. 75°03 75°2
Dispersion equivalent ......... 4°55 4 84

27 2

616 GLADSTONE ON ESSENTIAL OILS.

The coincidence of the calculated numbers with the results of expe-
riment is not so close as in the case of the terpenes, but it leaves little
room for doubt.

The hydrochlorides of the citrenes seem always to contain 2 mols.
HCl, which, with other considerations, has generally led chemists to
come to the same conclusion as to their constitution as that deduced
above.

Allied to the citrenes are several hydrocarbons produced from
essential oils, but having a different amount of hydrogen. If CyHy
has two pairs of carbon-atoms doubly-linked, C,)H;, ought to have only
one, and C,)H») ought to be a saturated hydrocarbon, while on the other
hand C,H, ought to have three. All this is actually the case.

CioHx is a hydrocarbon, which, according to Armstrong and Tilden,
frequently occurs among the products of decomposition of the essen-
tial oils (Trans., 1879, 745), and which Armstrong considers to
be a saturated body, and now designates “ cymhydrene.”’

C,,H,s is the menthene derived by Atkinson and Yoshida (Trans.,
1882, 49) from oil of peppermint.

CH, is cymene, which is easily prepared from the essential oils
by chemical reactions (Wright, Chem. Soc. J., 187, 6863), and
probably pre-exists in some of them. It is a well-known member
of the aromatic group of hydrocarbons. The figures given below
are the mean of 16 different preparations of Wright, described in
British Association Reports, 1873, 217, and 1874, 20.

That these form a true series is evident from the following table :—

tienen | | Boiling Temp.| Specific | Refractive | Specific Specific
| point. C. | gravity. index. | refraction. | dispersion.
| |
CyHen....| 170°? | 12° | 09-8046 14369 0°5430 | 0°0246
C..H,.....| 167-4] 8-5 | 0-8137 1°4460* | 0°5481 0 0313
eee — | 0°8483 1 -4677 | 0°5513 0 0334
CoHy.++-| 178 | 15 | 0 °3605 1-4818 0°5601 0 -0406

The refraction and dispersion equivalents of these for A and H is
given below, together with the theoretical values calculated on the
supposition that the four substances are of the constitution indicated

above.

* Reduced from C to A.

GLADSTONE ON ESSENTIAL OILS. 617

The Cedrenes.

Some of the essential oils contain hydrocarbons of the same empi-
rical composition as the terpenes or citrenes, but differing from them
considerably in physical properties. They are less limpid at the
ordinary temperature, the specific gravity and boiling point are both
higher, they do not expand so much when heated, and are sparingly
soluble in alcohol. Their vapour-density points to the molecular
formula C,;H. The average properties of nine specimens in the
original list, including cloves, rosewood, cubebs, calamus, cascarilla,
and patchouli, all fairly accordant, are as follows. As the hydro-
carbon from patchouli seems to be a very pure cedrene, and I happen
to have examined three different specimens, the mean of them is given
in the second column :—

| All. | Patchouli.
a
Boiling point . oveces | 249° to 260° 54° to 260°
Specific gravity at 20°C. ... 0... 0°9173 0°9251
Refractive index for A. reve 1°4944 1°4979
DisperMhOR. «oc cece ccccce-cecee| 0°0293 0 ‘0270
Specific refractive energy....... | 0 °5390 0°5384
0 -0319 0 °0292

Specific dispersive energy.......

These figures agree fairly well with a specimen of cedrene from
cedar, more recently determined (Trans., 1884, 248), viz., specific
refractive energy 0°5377, and specific dispersive energy 0°0299. It
appears, therefore, that the cedrenes closely resemble the terpenes
in their optical properties, and they must be considered polymeric
with that group. None of the specimens examined were in fact as
refractive as the citrenes.

The following will be the equivalents calculated out, and compared
with polymerised terpene :—

* Kanonnikoff determined the refraction equivalent of his “ tetrahydroterpene,”’
CoH, at 76°6 for the line C, which corresponds well with the above for the line A.

GLADSTONE ON ESSENTIAL OILS.

Refraction equivalent. | Dispersion equivalent.

Found. Calculated. Found. Calculated.

Old observations (mean)
From patchouli
From cedar

To this view it may be objected that no graphic formula can be
framed for C,;H., containing 1} pair of carbon-atoms doubly-linked.
It must be remembered, however, that our numerical formule are
founded on the volume in the gaseous condition, and that we really
know nothing of the molecular volume of liquids. All that we can
deduce from the experiment is that the liquid terpenes and cedrenes
are polymeric in the strictest sense of the term, but what multiple
of CjoH,, either of them may be is beyond our conjecture at present.

Caoutchene.

Among the juices of plants which are supposed to contain a sub-
stance of the composition C,H, is caoutchouc, and so peculiar are
the qualities of this substance that it seemed desirable to examine, if
possible, its optical properties. Unfortunately I have not yet suc-
ceeded in isolating a hydrocarbon which can be depended on as being
free from oxygen, and pure, but several preparations have been made
from good white caoutchouc by soaking it for some time in chloro-
form and fractionally precipitating the dissolved portion by means of
alcohol. The first fractions thrown down were a viscid substance,
which, when dissolved in benzene, gives a specific refractive energy
which varies rather widely in different specimens, but averages about
0°55. The specitic dispersive energy is about 0°037, and this suggests
the presence of two pairs of doubly-linked carbons. Indeed there is
some evidence of the existence of a substance having a still higher
dispersive power.

Caoutchene, the liquid CyH, obtained by the dry distillation of
india-rubber, has a boiling point and specific gravity similar to those
of citrene. A specimen that boiled between 172° and 178° yielded the
numbers given in the Appendix, from which may be deduced the
specific refractive energy of 0°554, and the specific dispersive energy
of 0:0366, giving—

Observed. Calculated.
Refraction equivalent 75°2
Dispersion equivalent 4°84.

GLADSTONE ON ESSENTIAL OILS. 619

In these optical properties, therefore, caoutchene also appears to be

a citrene.

Isoprene.

It is interesting to compare with these hydrocarbons another
liquid, derived from similar sources, having the same composition,
bat a smaller molecular weight. Through the kindness of Dr.
Tilden, [ have examined his specimens of pentine, derived by heat
from turpentine, and isoprene, by the dry distillation from caoutchouc,
both of which have the formula C;H,. The observations are given in
the Appendix, from which the following data are calculated :—

Sp. refractive Sp. dispersive
energy. energy.

0°5922 0°0479
0°5922 00460

This practical identity in optical properties confirms Dr. Tilden’s
opinion (Trans., 1884, 410) that the two substances are the same,
and the figures show them to be entirely distinct from both terpenes
and citrenes. It is necessary to conclude that no less than four of
the five carbon-atoms are doubly-linked, which would give—

Found. Calculated.
Refraction equivalent 40°27 39°8
Dispersion equivalent 3°22

This is in accordance with the conclusion of Tilden based on the fact
that they combine with four atoms of bromine, or 2 mols. HCI.

Chemical Structure.

If it be true that the amount of carbon doubly-linked varies in
different members of this isomeric group of hydrocarbons, it goes far
to explain the divergence in their properties. The differences in
their molecular structure must indeed be very important. Isoprene
would have to be expressed in a graphic formula by an open chain,
while the citrenes and caoutchene must be represented by means of
a closed chain; and terpene and camphene by some kind of double
ring. The great readiness with which cymene may be derived from
the various C, H;.’s, has led chemists generally to believe that these
compounds are allied to the aromatic compounds; this is confirmed
by their optical properties, and the formula suggested by Kekulé,
Fig. 1, may be retained for the citrenes, but it must be modified for
the camphenes and terpenes. The most symmetrical arrangement
would be that of Fig. 2; but Armstrong and Kanonnikoff have

620 GLADSTONE ON ESSENTIAL OILS.

already suggested graphic formule which are intended to express
the chemical reactions more fully, and either of which would meet
the requirements of the optical observations.

Fie. 1. Fie. 2.

C-0;H,
an
co? \on,

cH CH,
N,
O-CHs, oH, cif,

Another cause of difference in these and isomeric compounds may
possibly be indicated by a peculiarity of their dispersion equivalents,
It is found, as already stated, that the double linking, in the allyl
compounds for instance, affects the dispersion to a much smaller
degree than in the aromatic bodies, and it may eventually prove that
there is a similar cause for the smaller dispersion observed in the
case of camphene and the citrenes. This points to some difference of
structure hitherto unrecognized ; but it can as yet be viewed only as
a speculation.

Other Optical Evidence.

The phenomena of refraction and dispersion are not the only ways
in which light may reveal to us the structure of these hydrocarbons.
The phenomena of absorption also give us indications. Thus Hartley
shows (Proc. Roy. Soc., 1879, 290), that terpenes and citrenes
absorb the ultra-violet rays to a considerable extent, that cedrenes
absorb them to a still greater degree, but that neither they nor their
hydrates and oxides show the phenomena of selective absorption
characteristic of the aromatic group which contain three pairs of
carbon-atoms doubly-linked, and which is strongly marked wherever
cymene is present.

Abney and Festing (Phil. Trans., 1881, Plate 88) have examined
the absorption of the infra-red part of the spectrum by some of the
same substances, and here again the peculiar selective absorption
which characterises the members of the benzene-group is not shown
by oi] of turpentine. But there are other absorption-bands ; and
Captain Abney informs me that judging from his observations on
several of the essential oils and their congeners, he would consider
that they are rather of the nature of the aromatic than of the fatty
acid compounds.

GLADSTONE ON ESSENTIAL OILS.

Oxygen Compounds,

I had intended discussing the oxygen compounds in a similar
manner; but the subject presents great difficulties, arising partly
from the difficulty of obtaining these compounds in a sufficiently
pure condition, and partly from the double function of oxygen. Thus,
to take « special instance. Citronellol probably has the formula
Cy»H,,0, but the separation of it from other compounds contained in
the crude oil is extremely difficult. It is evident also that the atom
of oxygen may occupy the position either of that in alcohol or
of that in acetone. In the one case there would be one pair of
doubly-linked carbon-atoms more than in the other. The matter
is complicated by the fact that the oxygen would not have the same
refraction equivalent in the two cases. Where joined by single bonds
it has the value 2°, and where joined by double bonds 3°4, according
to Briihl. I find a similar difference in the dispersion equivalent, the
respective values being very wide apart, viz., about 0°10 and 0°18. 1
may return to this subject some day; in the meantime I will con-
teut myself with stating two or three conclusions that may be looked
upon as fairly established.

lst. Camphor, although its composition is CjH 0, has the optical
properties of a saturated compound, thus confirming the view which
has been entertained from its chemical properties. This suggests
its close relationship to the terpenes.

2nd. Besides camphor (CHO), cajeputol and its isomerides
(CjoH,,0), and menthol (C,H 0), have the refractive and dispersive
energy of saturated compounds, This gives much room for specula-
tion as to the differences between their constitutional formule.

3ed. Conversely, isomeric compounds containing oxygen may have
a different refractive energy. Thus terpinol (C;.H,,O), unlike caje-
putol, is shown both by its refractive and dispersive energy to have
one pair of doubly-linked carbon-atoms. Or to take C,H,,O, carvol
appears to have two pairs of doubly-linked carbon-atoms, while
thymol has certainly three.

4th. Anethol (C,oH,,0), belongs to another category. Its refractive
and dispersive energy are both too great to be explained by the
supposition of four pairs of carbon-atoms doubly-linked.

Since the foregoing was drafted, Kanonnikoff has sent me a copy
of his paper in the Journal fiir praktische Chemie, 33, 497, in which
I find that he discusses several of these oxygen compounds from
the point of view of their refraction equivalents, and has arrived
mainly at the same results.

GLADSTONE ON ESSENTIAL OILS.

Summary.

The preceding discussion may be summarised as to its main results
in the following table. Instead of repeating the refraction and dis-
persion equivalents, I have preferred returning to the specific refrac-
tive and dispersive energy. These numbers represent experimental
facts whatever the theory may be about the substance examined,
independent even of its composition. They are given in columns 2
and 3, while the figures deduced from my theory in regard to each of
the hydrocarbons are given in columns 5 and 6. That theory is
expressed in column 4, which contains the number of pairs of carbon-
atoms that I take to be in a “ doubly-linked” condition, on the basis
of each substance containing 10 atoms of carbon. Whether in fact
it is Cy, or Cys, or Cs, is not indicated by these optical properties, but
rests solely on other evidence. In estimating the value of the figures,
it must be remembered that the physical observations are trustworthy
to the third, if not the fourth, place of decimals, those for dispersive
energy being somewhat more exact than those for refractive energy ;
but the same amount of accuracy cannot of course be attained in
respect of the chemical purity of the different substances.

energy. energy. energy. energy.

‘543 | «=(0°0246 *543 0 -0243

Cymhydrene .......++-
Menthene

The terpenes ......+++.
Terebenthene.
Camphene........++-+
The cedrenes ......00+-
Ee GEONES 646 ccicccese
Isoterebenthene
eee

“548 | 0:°0313 *547 0 -0298
*537 0 °0295 537 0 0296
537 0°0294
528 | O
538 0°02
‘551L | O
552 0
“554 0
560
592

eoooocooooo

The advance in these experimental figures with the double linking
is evident enough, though it is obscured in the first two lines by the
larger amount of hydrogen contained in cymhydrene and menthene.
The influence of the double linking upon the specific dispersive
energy is apparent at a glance, and renders this property more
valuable than the refractive energy in determining the rational con-

GLADSTONE ON ESSENTIAL OILS. 623

stitution; though of course it is in the agreement of the two that we
may seek the most conclusive evidence.

APPENDIX.

The indices made use of in this paper, and not previously published,
are given in the following table. The specimens of pentine, isoprene,
and terpinol were kindly lent me by Professor Tilden; the sage-
terpene by Mr. Pattison-Muir, and the artificial anethol by Dr.
Perkin. Fresh determinations of cymene, carvol, and anethol are
added.

Refractive indices.
Specific
gravity.

Substance.
p-

“4742
"4667
“4696
"4835
*4750
*4079
“4041
*4838
*5020

Terpene from pine...... ; "8711

- SABC... ..0. ; "8632

“ er . * 8646
Cymene....ccccccceoce| 3 *8551
Caoutchene......eesee. ) 8449
ee : ‘6766
Isoprene cece *6709
Terpinol <aeneen *9296
COE cc ancck es eneesene! ‘9667
ee ‘9869
artificial ....... ‘9870

Ff tt et et et et
hn
fe et tt et et et

5614

”

The refraction and dispersion equivalents of the oxygen compounds
in the above table, together with a new determination of thymol dis-
solved in alcohol, are :—

Terpinol .... — 79°02 and 4°23
GEE 6.0600040+0 60% 7668 ,, 5°53
Thymol 78°91 ,, 6°38
Anethol (natural) .... 81°93 ,, 10°56

- (artificial).... 82°08 ,, 10°49

LVIII.—A Method of Investigating the Constitution of Azo- and
Diazo-derivatives and Analogous Compounds.

By Rapnart Metpona, F.R.S., Professor of Chemistry, and F. W.
STREATFEILD, Demonstrator of Chemistry in the Finsbury Technical
College, City and Guilds of London Institute.

In former papers by one of the authors (Meldola, Trans., 1883,
436, and 1884, 117), it was pointed out that the compounds re-
sulting from the action of diazo-salts of para- and meta-nitraniline
on 8-naphthylamine possessed neither the characters of true azo- nor
diazo-compounds, and it was suggested that these substances might
be related to the class of azimido-compounds of which the simplest
representative is Griess’s azimidobenzene :—

According to this view, the general formula of these compounds
may be written—

NH
X<yH>YY,

where X and Y may be similar or dissimilar, and the compounds may
be regarded as derivatives of the typical hydrazimido-compounds :—

NH
X<yy>NE.

In the paper referred to, it was further suggested that the difference
in the constitution of these 8-naphthylamine compounds resulted from
the interaction of the azo- and amido-groups, which in all probability
occupy an ortho-position with respect to one another, so that the
investigation of the supposed azo-compounds derived from orth-
amido-compounds becomes a matter of considerable importance. The
necessity of continuing the investigation in this direction has always
been kept in view, although, since the publication of the paper in
question, much additional evidence has been gained by the researches
of Zincke and his pupils (Ber., 1885, 3132 and 3142; Lawson, ibid.,
796 and 2422, and Inaug. Diss., Marburg, 1885; Sachs, Ber., 1885,
3125).

Among the recent results, the most important in relation to the
constitution of the compounds under consideration is undoubtedly the
discovery that they can be oxidised to substances of the azimido type

(Zincke, Ber., 1885, 3134). Upon this evidence Zincke accepts the

CONSTITUTION OF AZO- AND DIAZO-DERIVATIVES. 625

conclusion which had been arrived at on other grounds in the paper
published in the Transactions of this Society in 1884, viz., that the
compounds in question no longer contain an amido-group, but two
NH-groups. With respect to the constitution of these compounds

Zincke gives the two formule-—
x—n<NHyy x—NH. | SY
—_ <NH : \y’ ’
1. II.

the first of these corresponding with that proposed by one of the
authors of the present paper in 1884, and the second corresponding
with the formation of 8-naphthaquinone-hydrazide from §-naphtha-
quinone and phenylhydrazine as observed by Zincke and Bindewald

(Ber., 1884, 3026) :—

O(a
C,,H,O, + H.N-NH:C,H; = Cul | NH C Wl + OH:,.
‘ “Usils

The evidence at present available does not render it possible to
decide between these formule, and further investigation therefore

appears necessary.
In a note published by one of the authors (Meldola, Chem. News,

Dec. 5, 1884), it was suggested among other methods of investigating the

constitution of these and analogous compounds, that the behaviour of
the alkyl-derivatives, on reduction, might throw light on the question.
Thus, with respect to the two formule given, a monalkyl-derivative of
I would give, on complete reduction, X‘NH, and NH‘Y-:NR; and a
monalkyl-derivative of II would give either X‘NHR and NH,Y-NH,
or X:NH, and NH,*Y-NR, according to the position of the radicle.
A dialkyl-derivative of I would give on reduction X*-NH, and
NR-Y-NR, and of II, X*-NHR, and NH,Y-NHR. The same con-
siderations evidently apply to diazoamido- and amidoazo-compounds,
and, as far as we are aware, the decomposition products of the
alkyl-derivatives of these compounds have not hitherto been investi-
gated. Thus mono- and di-alkyl-derivatives of true amido-azo-
compounds of the type X-N.,"Y-NH, would give on reduction X-NH,
and NH,-Y-NHR, or NH,Y-NR,, whilst diazoamido-compounds of the
type X:N,,NH-‘Y could only give monalkyl-derivatives, X°N,NR-Y,
which, on reduction, would yield X-NH, and Y-NHR.

It appeared also that the present method would be capable of
determining finally the question of the symmetrical or asymmetrical
formula of the diazoamido-compounds. The symmetrical formula for

diazobenzeneanilide—

N-C,H;
HNC | ;
N: C.H;

626 MELDOLA AND STREATFEILD: THE CONSTITUTION OF

is given by Victor Meyer (Ber., 1881, 2447), but is rejected by him
as having little probability. The same formula has recently been
brought forward by Friswell and Green (Trans., 1885, 924), who
state that it appears to them to be more probable than the asym-
metrical formula. Supposing the diazo-amido-compounds to have
the symmetrical formula above given, it is evident that their monalky]l-
derivatives on reduction would give X-NH,.,Y-NH, and NRA).

In order to apply the present method, ethyl iodide was used as the
exploring agent, and, although we have as yet only studied one
compound, the results are sufficiently interesting to warrant their
being communicated to the Society.

We propose to extend the investigation to other diazo-, azo-,
azimido-, and oxyazo-compounds.

Paradinitrodiazoamidobenzene.

In the present communication, we give the results of the application
of the method indicated to the above-named substance, which has not
hitherto been prepared, the only analogous compound at present
known being the supposed dinitrodiazoamidobenzene obtained by
Hallman by the action of nitrous acid on metanitraniline (Ber., 1876,
390). We propose to apply our method to Hallman’s compound at
an early opportunity. In starting from paranitraniline for our first
experiments, we had two objects in view.

In the first place it appeared probable that the compounds resulting
from the action of nitrous acid on the nitranilines would be of a
distinctly acid character owing to the presence of the nitro-groups,
thus offering the possibility of easy displacement of H by alkyl
radicles in NH-groups which might be present, since the acidity of
the molecule rendered it probable that the H in the NH-group would
be displaceable by metals, and the salts thus formed would be decom-
posable by alkyl iodides, &c. In the next place, in the case of
paranitraniline, where the para-position with respect to the NH,-
group is occupied, it appeared probable that a true diazo-compound
would in the first place be formed, and that if this could be trans-
formed into its isomeric amido-azo-compound the N,-group would
occupy the ortho-position with respect to the amido-group. Thus
there might be formed one of the compounds represented by the
following formule :—

NO,C,H, N,N H-C,H,NO,, NO,°C,HyN : N-C,H;(NO,)-NH,,
(4) (1) 7 (4) (4) oe (4) (1)

> NH HN.
NO,C,H,N <nH> Ces NO,, NOyCoHy NH | os NO.

III. IV.

[FF YY eee

AZO- AND DIAZO-DERIVATIVES. 627

As the result of our experiments, we find that with paranitraniline
a true diazo-amido-compound (formula I) is alone produced.

In order to prepare this substance, paranitraniline was dissolved in
the minimum quantity of strong hydrochloric acid, and the cold
solution diluted with water, but not so much as to throw out the
nitraniline. The solution of sodium nitrite (1 mol. to 2 mols. nitr-
aniline) was then gradually added, and the mixed solutions allowed to
remain in a cool place for 12—14 hours. The separation of the
diazo-compound commenced when the whole of the nitrite had been
added, and at the expiration of the time mentioned the contents of
the flask had become semi-solid. The pulpy yellow precipitate was
collected, washed thoroughly with cold water, and a specimen
purified for analysis by crystallisation from alcohol. It forms
small yellow needles soluble in boiling alcohol and acetone, but not
very readily, and only sparingly soluble in chloroform and the benzene
hydrocarbons. It melts at 223°, at the same time frothing up and
decomposing. The following results were obtained on analysis :—

I. 0°1346 gram gave 0°2489 gram CO, and 0°0372 gram water.
II. 0°0843 gram burnt in a. vacuum with CuO gave 17°8 c.c. N at
15°5°, and 760°1 mm. bar.
Theory for

NO,-C,Hy: No NH-CyHy-NOg. Found.

50°43

3°07

24°66

The substance, as anticipated, possesses distinctly acid properties,
dissolving in cold alcoholic potash or soda with a splendid magenta-
red colour, and in boiling aqueous alkalis with a similar colour. The
alkaline solutions show a remarkable stability, the substance forming
well characterised crystalline salts which separate out from the solu-
tions on cooling. The sodium salt, which probably has the formula
NO,C,HyNaNNa‘C,H,NO,, forms beautiful steel-blue needles, but
these are stable only in the presence of excess of alkali; when collected
and washed, they gradually decompose on exposure to the air with the
liberation of the original yellow substance. The acidity of the
substance is sufficiently great to decompose alkaline carbonates, the
characteristic red colour being developed on boiling the diazo-com-
pound with a solution of potassium or sodium carbonate. The
stability of the substance in the presence of alkalis was shown by
boiling some of it for two days with an excess of strong caustic potash
solution; the crystalline potassium salt, on being decomposed by
dilute sulphuric acid, gave the original substance unaltered. The
cold alkaline solution of the compound, when acidified, gives the

628 MELDOLA AND STREATFEILD: THE CONSTITUTION OF
original substance in the form of a greenish gelatinous precipitate
which gradually becomes yellow on standing; on acidifying the hot
alkaline solution, the substance immediately separates in a yellow
flocculent state. It is very unstable in the presence of acids, being
decomposed by boiling with glacial acetic acid with evolution of
nitrogen and the formation of resinous uncrystallisable products.
Heated with dilute sulphuric (1 part strong acid to 4 of water) in a
sealed tube for 2—3 hours at 100°, the substance is decomposed with
the liberation of nitrogen and the formation of paranitraniline and
resinous products. No nitrophenol appears to be formed in this
reaction. A similar decomposition was effected by heating the
substance with water in a sealed tube at 230° for three hours, the
product being, as before, paranitraniline and resin.

Products of Reduction.

Supposing the substance to have been an amidoazo-compound, it
seemed advisable to study in the first place the action of mild reduc-
ing agents in order to see whether a hydrazo-compound could be
obtained. By the action of ammonium sulphide on a warm alcoholic
solution of the sodium salt, reduction was effected and the magenta-
red colour of the solution gradually changed to orange. On dilution
with water, a reddish flocculent substance separated out, but this did not
appear to be a very definite compound, as it could not be crystallised,
and had no definite melting point. It decomposes when heated, and
a specimen that had been placed to dry in the water-oven suddenly
frothed up and decomposed with carbonisation. The dry substance
forms a reddish powder which deflagrates vigorously on ignition.
We have since found that more satisfactory results are obtained by
reducing the hot aqueous solution of the sodium salt with ammonium
sulphide. When the red colour of the solution has disappeared,
reddish-brown needle-shaped crystals are deposited on cooling. We
propose to make a further study of this reduction product, but the
main question of the constitution of the original substance is for the
present settled by the fact that on complete reduction it gives only
paraphenylenediamine. This was proved by a careful examination of
the reduction products obtained by the action of zinc-dust and acetic
acid and of tin and hydrochloric acid on the compound dissolved
in alcohol. In both cases, paraphenylenediamine was the sole product,
and was identified by its characteristic colour reactions with ferric
chloride, potassium dichomate, ferric chloride in the presence of
hydrogen sulphide, and by the analysis of the diacetyl-derivative.
The formule IT, III, and IV are thus excluded, as these substances on
reduction would give a mixture of di- and tri-amidobenzenes, and the
substance is therefore a true diazoamido-compound.

AZO- AND DIAZO-DERIVATIVES.

Decomposition by Hydrochloric Acid.

When decomposed by heating with acids‘in the presence of water,
diazoamido-compounds usually break up into amido-derivatives,
phenols, and nitrogen. It appears, however, from Lawson's ‘experi-
ments (Ber., 1885, 797) that this mode of decomposition is not
confined to diazoamido-compounds, since the compounds obtained by
the action of diazotised amines on §-naplithylamine also yield the amine
(8-C,.H;NH;) and the corresponding phenol (C,H;-OH,NO;C,H;OH,
&c.). That these B-naphthylamine compounds are not true diazo-
amido-compounds is shown by the fact that they give naphthylene-
diamine on reduction.* The failure to obtain nitrophenol from the
present dinitrodiazoamidobenzene by decomposing it with dilute
sulphuric acid is thus a feature in the character of the diazoamido-
compounds to which it seems desirable to call special attention. Much
more definite results were obtained by decomposing the substance
with strong hydrochloric acid, three hours’ heating in a sealed tube in
a water-bath being sufficient for complete decomposition. Nitrogen is
given off on opening the tube, the contents consisting of a mixture of
large needle-shaped crystals and a dark tarry substance, together
with the excess of acid liquid. The crystals were identified as the
hydrochloride of paranitraniline, and the tarry substance, after being
repeatedly washed with warm dilute hydrochloric acid to completely
remove the paranitraniline, was found to be separable by steam distil-

lation into a white crystalline substance having an odour of bitter
almond oil, and a small quantity of a dark oil which remained behind
in the flask and solidified on cooling to a brittle resin. The crystalline
distillate was purified by redistillation in a current of steam, and
proved to be paranitruchlorobenzene : m. p. 83° :—

0°1737 gram burnt with lime gave 0°1590 gram AgCl.

Theory for
C,H,-Cl- NO). Found.
> aa 22°54 22°63

The decomposition of the diazoamido-compound by hydrochloric
acid may thus be simply represented by the equation—

NO,C,H,y} »NH:-C,HyNO, + HCl = NO.*C,H,Cl os N, a
NH,-C,H,NO,,.

It seems most probable that the nitraniline residue to the left of
the N,-group in the above formula, that is the nucleus directly

* According to a recent paper by Nietzki and Goll (Ber., 1886, 1281), amido-
az0-8-naplithalene can be diazotised under suitable conditions, and thus appears to
contain an NH,-group.

VOL. XLIX. 2uU

630 MELDOLA AND STREATFEILD: THE CONSTITUTION OF

attached to the azo-group, is the one into which the chlorine of the acid
enters to displace the N;.

This point is, however, sufficiently interesting to examine into
further, as it is by no means certain that intramolecular rearrange-
ment may not occur during the decomposition of diazo-compounds,
owing to the formation of intermediate additive products. We
propose to extend our investigations in this direction by examining
the decomposition products of mixed diazoamido-compounds contain-
ing dissimilar radicles.

Paradinitrodiazoethylamidobenzene.

In order to prepare this compound, the dinitrodiazoamidobenzene
was dissolved in alcohol with the addition of the theoretical quantity
of caustic potash to form the potassium salt. The necessary quantity
of ethyl iodide was then added and the liquid kept boiling in a flask
with an upright condenser for about 10 hours, when the red colour of
the solution had disappeared and a heavy yellowish crystalline
deposit had separated: 20 grams of the dinitrodiazo-compound gave
about 18 grams of the ethyl-derivative, or about 81 per cent. of the
theoretical quantity. The ethyl-derivative was collected, washed
with water to remove potassium iodide, and crystallised twice from
alcohol in which it is but very sparingly soluble. The substance
forms small yellow needles, melting without decomposition at 191—
192°.

I, 0°2825 gram gave 0°5501 gram CO, and 0°1072 gram OH2.
IT. 01456 gram gave 0°2845 gram CO, and 0°0550 gram OH).

II. 0°0958 gram burnt into vacuum with CuO gave 17:9 c.c. N at

13:7° and 757 mm. bar.

Found.
Theory for vom os —~
NO,-CgH,:No:N (C.H;)-CgsH4-NOg. q IT. ITT.
53°33 SS 43°28 —
419 —
_- 21°9

The ethyl-derivative is no longer soluble in alkalis, thus confirming
the view that the salt-forming power of the origina] substance is due
to the hydrogen of the -N,,NH- group. The compound is quite as
unstable towards acids as the original substance, being decomposed
by glacial acetic acid, dilute sulphuric acid, &c., but, as before, no
nitrophenol is formed. The decomposition with hydrochloric acid
was effected by heating in a sealed tube in a water-bath for 2 to 3
hours. The products of this reaction proved to be naranitrochloro-
benzene and ethylparanitraniline :—

AZO- AND DIAZO-DERIVATIVES. 631

NO,C,Hy NN (C.Hs)°C.Hy NO, + HCl = NO,°C,H,Cl + N, +
NO,C,HyNH-C,H;.

The last-named compound, which was first described by Weller
(Ber., 1883, 31), was identified by its melting point (95—96°) and
crystalline form, as well as by the analysis of its nitroso-derivative
which has not been hitherto described.

Ethylparanitropheny lnitrosamine, NO."C,Hy N (NO)-C.H;.—On adding
the necessary quantity of sodium nitrite dissolved in water to the
hydrochloric acid solution of ethylparanitraniline, a white flocculent
precipitate of the nitrosamine at once separates. This was collected,
washed with water, and crystallised from alcohol. It forms large
straw-coloured needles melting at 119°5°. The corresponding ethyl-
metanitrophenylnitrosamine melts at 47° (Nélting and Stricker, Ber.,
1886, 546).

The decomposition of the ethyl-derivative by hydrochloric acid into
ethylnitraniline and nitrochlorobenzene is conclusive evidence against
the symmetrical formula—

N-C,H, NO,
CHNC l

N-C.H,-NO,

We have not yet applied the present method of investigation to the
simpler unsubstituted diazoamido-compounds, but we have no doubt
that it will be found generally applicable. Alkyl-derivatives of such
diazoamido-compounds have however been prepared by Gastiger
(Bull. Soc. Chim., 42, 338) and by Nélting and Binder (ibid., 42,
336 and 341), by acting with diazobenzene and paradiazotoluene
chloride on ethylaniline. The resulting compounds have the formule
C;H;*N2*N(C.H;)-C.Hs, CHs*CsHy NN (C.H;)*C.Hs, &c., and behave as
true diazoamido-compounds when decomposed by acids, the products
being phenol and ethylaniline, and paracresol and ethylaniline
respectively.

As the results of this investigation showed that the substance
obtained by the action of nitrous acid on paranitraniline is a true
diazoamido-compound, an attempt was made to convert it into its
isomeric amidoazo-compound by dissolving it in fused paranitraniline
containing paranitraniline hydrochloride. Under these circumstances,
however, complete decomposition takes place with the evolution of
nitrogen and the formation of an uncrystallisable resinous substance.

632

LIX.—The Formation .and Destruction of Nitrates and Nitrites in
Artificial Solutions and in River and Well Waters.

By J. H. M. Munro, D.Sc., College of Agriculture, Downton,

Salisbury.

CONTENTS.
PAGE
Introduction ......seeeeeeeee: cvcccccccccccccccccccccesces «©6683
I. Nitrification of éthjlamine by soil. . Vttetreeeeneeteceencs Ti
II. ” potassium thiocyanate by BSlhcccccccccccceccsecee GBF
III. ” ammonium thiocyanate by soil..........+..-++++++ 638
oc agapenamaamaa cenesseccessceccsssocscvvccs 6D
IV. Nitrification of uren ......cccccecccecccccereccccccccccesccsess 639
¥, - of gelatin by soil. ecceeee coccccccces§ 6G
VI. P of urine in accidentally fertilised selutlons. pececcescees 642
VII. a of ammonium salts .......... 643

Nitrification in strong-and weak solutions fs ammonium chloride 643
in unfertilised solutions of ammonium chloride,

”

sterilised by heat, and not sterilised ........... 645

- of ammonium oxalate ; decomposition of oxalates by
soil ferments......... 645

” of ammonium chleride wttheu purposely “added
organic-carbon........ +. cccccccese CG
” in boiled, ‘filtered, wansporent atin (co 648
VIII. Is organic carbon essential to nitrification? ...........0..e.ee00ee. 651
IX. Nitrification of ammonium salts in natural waters. cece -ee 654
Experiments on the nitrifying power of river and ‘well wate. . 658
X. Denitrification, and the influence of organic matter on nitéifeation. . 667
Alternation of nitrification and denitrification ...........0++0+ 669

XI. Period of incubation of the nitrifying organism.............+.+.+. 679
BUBOMTEE cc cccccccccccecccccccccocccccsccceccessccocceccece §€=©6GBL

Ix the spring of 1883, attracted by MM. Schloesing and Miintz’s
account of the nitrifying organism of ‘the soil, and by Mr. R.
Warington’s extremely interesting researches-on “ Nitrification,” which
had been published in the Journal of the Chemical Society (Trans.,
1878, 44, and Trans., 1879, 429;), I commenced a few observations on
the same subject. My first series of experiments was commenced
before I became acquainted with Mr. Warington, and indeed, without
the knowledge that he was pursuing his investigations on this subject.
This series was designed to repeat and corroborate some of Mr.
Warington’s observations with ammonium salts, and to extend the
results to three or four other nitrogenous compounds. Later on, I com-
municated with that gentleman, and received from him advice derived

MUNRO: FORMATION AND DESTRUCTION: OF’ NITRATES. 633

from his own experience, and encouragement to pursue my observa-
tions with the view of eliciting new facts.

So slow, however, is the nitrifying process under some conditions,
that this first series remained under observation for nearly two years,
whilst subsequent nitrifying solutions, started in the summer and
autumn of 1883, are only just done with. The great length of time
occupied in these experiments is perhaps the reason why so few
English chemists appear to: have been attracted to the subject.
Under these circumstances observations which corroborate the state-
ments of others have some value, and in accordance with Mr.
Warington’s wish, I now condense into a brief form such of my
results as appear of any interest, in view of the continuation of his
researches (Trans., 1884, 637), together with some experiments on
substances other than ammonium salts, on nitrification in natural
waters, and om denitrification, which I believe to be original. The
observations now described were made in 1883 and the spring of 1884,
that is, between the publication of Warington’s second (1879) and
third (1884) reports; the only exceptions are certain experiments
which I have made within the last two months, to elucidate points
which occurred to me during the ‘preparation of the paper for
publication.

To attain as much clearness as possible, 1 have arranged the ex-
periments in something like chronological order under separate
headings, and will consider in succession—

Nitrification of Netrogenous Compounds other than Ammonium Salts.

Nitrification of Ammonium Salts in Artificial Solutions.

Nitrification of Ammonium Salts in Well and River Waters.

Denitrification, and the Influence of Organic Matter on Nitrification.

I. Nitrification of Ethylamine by Soil.

We have direct and conclusive evidence that plants avail them-
selves of nitrogen supplied as nitrates, and that the nitrates are
absorbed as such by the plants, and in some cases stored up without
immediate alteration in their tissues. Taking into consideration with
this the fact that all organic manures evolve ammonia during decom-
position, and that ammonia in the soil is rapidly transformed into a
nitrate, it may well be doubted whether plants ever assimilate
nitrogen directly from any other compounds than nitrates. It was
for this reason that it seemed to me interesting to ascertain whether
ethylamine is nitrified by contact with soil, for G. Ville has long ago
shown (Recherches sur la Végétation, 1857) that ethylamine and am-
monia are equally efficacious as nitrogenous manures. Trimethy]-
amine has also been successfully tried as a manure in France.

634 MUNRO: THE FORMATION AND DESTRUCTION OF

The sample of ethylamine hydrochloride employed was tested for
purity by conversion into aurochloride. The crystals of ethyl-
ammonium aurochloride on ignition gave 50°78 per cent. metallic gold,
whereas the formula C,H,;-NH;Cl,AuC}, requires 51°04 per cent.

I experimented in all with three solutions of ethylamine hydro-
chloride. Two of these were made up on March 9, 1883, and the
description of the mode in which they were prepared and examined
applies to all the solutions of my first series, to which the same date
is affixed. The modifications adopted in later series are noticed in
their place.

309 mgrms. (Bl) and 618 mgrms. (B2) ethylamine hydrochloride
were dissolved in ammonia-free distilled water in two stoppered
bottles, to each of which I then added—

222 mgrms. potassium hydrogen tartrate,
222 »» magnesium sulphate, cryst.,
222 sodium phosphate, cryst.,

”

and water to 500 c.c. Half a teaspoonful of precipitated calcium
carbonate to render the mixture permanently alkaline, and 2 grams
moist arable soil, weve put in afterwards. All the solutions of this
date were kept in stoppered or corked bottles or flasks, about half

full, which were placed in a cupboard in the laboratory, and exposed
to the ordinary variations of temperature of the place. Chiefly on
account of the low temperature prevailing, nitrification did not take
place in any of the solutions of this series until the beginning of
June, but when once commenced, it proceeded in several cases with
great rapidity. The solutions were examined from time to time, for
ammonia by the Nessler reagent, for nitrites by permanganate, and
for nitrates or nitrites by the indigo test as used by Boussingault. It
was only in later experiments that I adopted the more delicate re-
agents, metaphenylenediamine and diphenylamine, in place of per-
manganate and indigo for qualitative testings. The quantitative
determinations were made as follows :—

Ammonia, by Nesslerising a diluted fraction of the liquid.

Nitrites, by weak standard permanganate (1 c.c. = 0°0001 gram
available oxygen) in a pipetted fraction of the liquid aciditied with
sulphuric acid.

Nitrate in absence of nitrite, by the indigo process as modified by
Warington.

Nitrate in presence of nitrite, by oxidising the nitrite in a pipetted
fraction by cautious addition of permanganate, then determining the
total nitrate with indigo, and deducting the nitrate formed from
nitrite.

These methods, which are the same as those used by Warington,

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 635

enable one to follow very fairly the progress of a nitrifying solution,
but they do not yield results of great accuracy. In some of my later
experiments, for example, I wished to ascertain whether the nitrifica-
tion was solely nitrous, or whether a small quantity of nitrate co-
existed with the nitrite, and this it was extremely difficult to do, if
not impossible. The estimations of ammoniacal nitrogen are merely
approximate in all cases, since any error made in Nesslerising the
very small fraction of liquid available for each testing is greatly
multiplied when caiculated on the whole 500 c.c.

The progress of the solutions Bl and B2 is recorded in the sum-
mary of observations made at the following dates :—

Bl.—Made up March 9, 1883. 500 c.c. contained 309 mgrms.
ethylamine hydrochloride = 53 mgrms. nitrogen. No ammonia,
nitrites, or nitrates, were detected up to March 26. Between this
and April 13, however, the solution became turbid, small patches of
mycelium began to appear throughout the liquid, and ammonia was
formed. By July 11 the ammonia present amounted to 16 mgrms.
nitrogen, and nitrate was present. On July 26 the N as nitrite
amounted to 28 mgrms., but between this date and August 16 it was
completely converted into nitrate. Nitrification was practically com-
plete on this date, and the N present as nitrate was estimated as
474 mgrms., or 90 per cent. of that originally taken.

B2.—Made up March 9, 1883, was double as strong as the pre-
ceding, and contained 106 mgrms. N in 500 c.c. Fora time this followed
a similar course to Bl. On July 11 the ammonia present was
equivalent to 33 mgrms. nitrogen, and the nitrous fermentation pro-
ceeded with such rapidity from this date, that by July 26 all the
nitrogen was present as nitrite (105 mgrms. estimated). Instead of
suffering conversion into nitrate by August 18, however, the nitrite
was quite unaltered on that date, the quantity of permanganate con-
sumed being exactly the same as before. Conversion into nitrate set
in some time during the long vacation, for on October 18 no nitrite
was present, and the N as nitrate was estimated as 104 mgrms., again
over 90 per cent. of that originally present as ethylamine.

These two experiments show that ethylamine is easily nitrified, and
that the process is divisible into three stages. At tirst ammonia is pro-
duced, and this may suffer total conversion into nitrite before any nitrate
is formed (B2) ; the nitrite may be rapidly converted into nitrate, or
may persist for a time without alteration (B2), ultimately, however,
undergoing complete oxidation. The apparently capricious persist-
ence of nitrite, ending sometimes in sudden and rapid conversion
into nitrate, occurred in many of my experiments, as well as in those
of Warington.

The third solution of ethylamine was made up on July 20, 1883,

636 MUNRO: THE FORMATION AND DESTRUCTION OF

and belongs to.a series of mixtures nearly all of which failed to nitrify
for 13 months, and then succeeded only after: re-seeding with soil.
At that date, I was anxious to obtain nearly clear solutions in which
to observe nitrification, and I therefore tried to replace the
carbonate of lime by soda or potash, and to substitute some other
form of organic matter for the alkaline tartrate hitherto generally
employed, and supposed to serve as aliment for the nitrifying or-
ganism. When an alkaline tartrate or any similar substance is
employed, a bacterial fermentation is set up almost immediately,
which destroys nitrates if they happen to exist already in the liquid
experimented on, and at any rate renders the whole liquid turbid
from the presence of myriads.of organisms. These die down after a
few days and form a deposit at the bottom. of the liquid, but they
never leave the liquid absolutely clear. At alater stage I was able to
avoid these bacteria altogether, to do without tartrates or similar
substances, and to obtain almost perfectly clear nitrifying solutions,
but my earliest attempts. almost all failed at the first seeding,
apparently because 1 or 2 mgrms. soil is insufficient to induce nitri-
fication in unfavourable solutions.

B3 contained no organic matter beyond the ethylamine itself, and
that in the milligram. or two of added soil. The solution was made

up to contain in 100 c.c. of ammonia-free distilled water—

3l mgrms. ethylamine hydrochloride (= 5°3 mgrms. N).
20 » MgS0O,, cryst.

20 » NaHPO,, cryst.

17°7 » KHO.

The alkalinity of this solution. is equal to about 145 mgrms. KHO
per litre (= 36°25 ammoniacal N per million), after allowing for
complete saturation of the phosphoric acid. The solution about half
filled a 6-oz. stoppered reagent bottle, and was seeded with 1 or
2 mgrms. only of dry soil from 3 or 4 inches depth. It remained quite
clear (except for a few flocks of magnesium phosphate) until February
2, 1884, and contained no ammonia, nitrite, or nitrate on that date.
On April 7, 1884, it was still clear, but a little ammonia was present,
and on August 21, 1884, more than a year after seeding, although
ammonia was still present, no nitrate or nitrite could be detected,
even with diphenylamine. Some time before this, viz., on Novem-
ber 10, 1883, thinking that the failure to nitrify might be due
to the exclusion of the alkaline tartrate, I had divided this solution
(together with others which had been made without tartrate, and had
refused to nitrify) into two portions, one of which I placed in an
open wide-mouthed bottle, with a paper cap. To this part of the
solution I added a small crystal of Rochelle salt, and the bottle was

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 637

kept in a cupboard in a warm room all the winter. Yet although
a delicate cloudlike fungus gradually: grew in the solution, no nitrite
or nitrate had developed. by April 7, 1884; the solution at this date
having almost dried up, was thrown. away.. I may mention here that
in no instance did the addition of Rochelle salt to the solutions which
had refused to. nitrify, produce the slightest formation of nitrite or
nitrate, although in every case it was speedily followed by various
growths of mould. Having thus come to the conclusion that
absence of the alkaline tartrate was not the- reason of non-nitrifica-
tion in the portion of solution which had been kept in the
stoppered bottle until August 21, 1884, I on that date re-seeded it with
a few more milligrams soil, and on November 13, 1884, I had the
satisfaction of finding it in.full nitrification, both nitrite and ammonia
being present in considerable quantity. This solution was unfavour-
able for nitrification, since it contained a caustic alkali, but the insuffi-
cient quantity of soil added at first was evidently the prohibitive
circumstance. It remained free from mould throughout, and is at
present nearly clear. The nitrite present on November 13, 1884,
has since been converted into nitrate, but ammonia is still present
(March, 1886). The original alkalinity of the solution, in fact
(= 145 mgrms. KHO), is.insufficient to allow of complete nitrification,
since the hydrochloric and nitric acids produced from 31 mgrms.

ethylamine hydrochloride require 42°4 mgrms. KHO for saturation.

A plain solution of ethylamine hydrochloride in distilled water, con-
taining 3 grams per litre, has remained perfectly free from ammonia
and nitrite from July 20, 1883, until new (March, 1886).

Il. Nitrification of Potassium Thiocyanate.

The thiocyanates, it is well known, are poisonous to plants, and it
occurred to me that possibly nitrogen in this form is not susceptible
of fermentation into nitrate. I therefore included potassium thio-
cyanate in my first series of solutions, made on March 9, 1883.
Solution (Cl) contained 243 mgrms. KCNS; 222 mgrms. potassium
bitartrate; 222 mgrms. MgSO, cryst.; and 222 mgrms. Na,HPO,
cryst., in 500 ¢c.c. ammonia-free distilled water. Calcium carbonate
and 2 grams moist soil were added. Total N= 35 mgrms. Until
March 26 no change occurred. Ammonia then slowly developed, and
a mould began to grow. On July 19, 33 mgrms. N existed as
ammonia, and a trace of nitrite was present. On August 16 the
ammonia had fallen to 12 mgrms. N, and the nitrite had increased ;
as the thiocyanate itself decolorises permanganate, no estimation
was possible. A little nitrite was still present on October 18,
although there was then no ammonia. On October 29, both nitrite

635 MUNRO: THE FORMATION AND DESTRUCTION OF

and ammonia were ahsent, and the N as nitrate was estimated at
27 mgrms., a nitrification of nearly 80 per cent. As showing the
complete oxidation of the thiocyanate, I may mention that when
made up, and even on March 26, 5 c.c. of the solution decolorised
14°4 c.c. of standard permanganate; on October 29, one drop of per-
manganate sufficed to colour the liquid. The reaction with ferric
salts disappeared before June 9.

It appears from this experiment that thiocyanates are easily
nitrified by soil, and that as in other cases, ammonia, nitrite, and
nitrate appear in succession. A plain dilute solution of potassium
thiocyanate in distilled water, made up March 9, 1883, is now, three
years later, quite free from nitrite and nitrate. It has grown a few
spots of mould, however, and contains a trace of ammonia.

III. Nitrification of Ammonium Thiocyanate. Experiment with
Thiocarbamide.

Although ammonium thiocyanate is a plant poison, its isomeride,
thiocarbamide, is stated by Dr. Emerson Reynolds to act as a nitro-
genous manure. It is curious that in the following experiments I
should have succeeded in nitrifying the former, whilst the latter
has resisted nitrification for nearly three years. The ammonium
thiocyanate resisted nitrification for 13 months, because of the
sterility of the 1 or 2 mgrms. of soil with which the solution was
seeded, but it commenced to nitrify very shortly after being re-seeded
with soil; re-seeding, however, has nut induced nitrification of the
thiocarbamide.

Ammonium Thiocyanate——Solution Gl, made up July 20, 1883.
The 100 c.c. ammonia-free distilled water contained 35 mgrms.
NH,CNS; 20 mgrms. MgSO, cryst.; 20 mgrms. Na,HPQ, cryst., and
2) mgrms. oxalic acid crystals neutralised by addition of 17°7 mgrms.
KHO. 1 or 2 mgrms. dry soil were added, and a little magnesium
carbonate. This bottle, it will be seen, belongs to the first series,
from which I excluded tartrates, on account of their liability to
encourage reducing bacteria and moulds. Organic carbon was
supplied in the form of an oxalate, and this made no difference
to the eventual nitrification of the solution, although I for some
time believed that it would. I now believe that the oxalate might
also have been omitted, but prefer to discuss the influence of organic
matter on nitrification in the light of subsequent experiments.

Progress of the solution:--Up to November 10, 1883, no change
could be detected, the solution giving the reactions for oxalates,
ammonia, and thiocyanates, remaining perfectly clear, and giving
no evidence of nitrite or nitrate with metaphenylenediamine or

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 639

indigo. Half of the solution was then removed to a shallow bottle,
a crystal of Rochelle salt added, and the bottle placed in a warm
cupboard for the winter. On February 2, 1884, it began to grow
a mould, which increased in amount until April 7, 1884, when
the solution was thrown away, having completely failed to nitrify,
notwithstanding the addition of tartrate. The half solution containing
no tartrate was kept in its stoppered bottle until August 14, 1884,
and as it showed no nitrification or other apparent change at this
date, 13 months after being made up, it was re-seeded with a few milli-
grams of soil. By November 13, 1884, abundance of nitrite was formed,
but ammonia was still present. Since then, both ammonia and
nitrite have been converted into nitrate, and the reaction with ferric
salts has disappeared. The addition of a sufficient quantity of soil
was therefore the indispensable condition of nitrification with this
solution.

Thiocarbamide.—I prepared this substance in the ordinary manner
by heating ammonium thiocyanate at 170° C. for some time, but I
could not free it entirely from traces of that salt. The solution
for nitrification was of the same strength, and was made up on the
same date (July 20, 1883) as the corresponding one of ammonium
thiocyanate, and it was in all respects similarly treated. I need
therefore only say that it entirely failed to nitrify, even after re-
seeding with soil, and on March 13, 1886, it contained neither nitrite
nor nitrate. The reaction with ferric salts, due to the small admixture
of thiocyanate, gradually disappeared, but the thiocarbamide appears
to have suffered no change.

Quite recently I have made a third attempt to nitrify this solution.
A gram of fresh soil was added three months ago, but as I write
(June 8, 1886) the solution is unaltered, and there is no sign of
nitrification. After three seedings, and a period of trial extend-
ing over nearly three years, we may conclude with tolerable cer-
tainty that thiocarbamide is not nitrifiable; very probably it exercises
an antiseptic action fatal to nitrification, for otherwise the traces
of ammonium thiocyanate with which it was contaminated would
certainly have furnished nitrate sufficient for detection.

Plain solutions of ammonium thiocyanate and of thiocarbamide
in distilled water (1 mgrm. per c.c.) have remained free from nitrite
or nitrate, and without apparent alteration, for nearly three years.

IV. Nitrification of Urea.

Diluted urine has been used as a medium for nitrification in many
of Mr. Warington’s experiments, and in the summer of 1883 I made
up a few solutions from instructions which he was kind enough to

640 MUNRO: THE FORMATION AND DESTRUCTION OF

giveme. I need only say that they nitrified quickly, passing through
the usual stages of ammonia, nitrite, nitrate. The two experiments
now to be described were made with pure crystallised urea, in order
to see whether nitrification would take place without the presence of
the organic matter which accompanies urea in urine. To the first
bottle (F1) I therefore added organic carbon in the shape of an
oxalate only ; the solution eventually nitrified (and I believe would
have done so without the oxalate). The second bottle (F2) was like
the first, but soil was entirely omitted. The behaviour of these two
solutions exhibits. very well the vital relation of soil to the process of
nitrification. F1 seeded at first with only a milligram or two of soil,
refused to nitrify for 13 months, when re-seeding with a few milli-
grams of soil speedily set up nitrification. F2, although not sterilised
by heat, has remained for 2 years. and 8. months without showing a
sign of nitrification. It developed ammonia, though more slowly than
F1; but in the absence of soil the ammonia showed no tendency to
pass into nitrite or nitrate, notwithstanding the supply of salifiable
base at hand. All the experiments have tended to convince me that
the ammoniacal ferment is far more generally diffused than the nitric
ferment.

The solutions were made up. July 20, 1883. Total N in each
= 9°3 mgrms.

Fl. 100 c.c. contained 20 mgrms. urea; 20:mgrms. MgSQ, cryst.;
20 mgrms. Na,HPO, cryst.;. and 20 mgrms. crystallised
oxalic acid neutralised with 17°7 mgams. KHO. A little
magnesium carbonate and 1 or 2 mgrms. soil were added.

F2. Exactly like Fl, but with no soil.

Aug. 1, 1883. Fl contains 0°2 mgrm. NH;. F2 contains no NH.

Aug. 16, 1883. Fl contains 0-4 mgrm. NH;. F2_,, ™

Aug. 21, 1883. Fl contains 0°'7 mgrm. NH;. F2__,, -

Oct. 18, 1883. F2 contains trace N H;.

Nov. 10, 1883. ¥1 contains 10 mgrm. NH. F2 contains 2°5 mgrms.

NH.

When once contaminated, therefore, F2 developed ammonia faster
than Fl. No nitrite or nitrate having appeared in either solution to
this date, half of each solution was removed to another bottle, and a
crystal of Rochelle salt added as explained in previous sections. These
tartrated solutions remained without nitrification (although the one
with soil eventually grew a mould) until April, 7, 1884, when they
were thrown away.

Aug. 14, 1884. Fl and F2 (not tartrated) are clear, free from
mould, contain ammonia but no nitrite or

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 641

nitrate. Fl was now re-seeded with a few
milligrams of soil.
Nov. 13, 1884. F1 contains ammonia and nitrite, and is therefore
nitrifying.
F2 contains ammonia but no nitrite or nitrate.
Mar. 13, 1886. F1 contains a trace of ammonia and much nitrite.
MgC, still in excess.
F2 contains much ammonia, no nitrite or nitrate.

The nitrification of Fl is, therefore, even now not complete, and is
chiefly, perhaps wholly, nitrous. The quantity of liquid remaining is
too small for analysis.

After the last testing, March 13, 1886, E2 was seeded with soil, and
it is now, June 10, 1886, in full nitrification.

A plain solution of urea in distilled water (5 mgrms. per c.c.)
behaved like F2. Soon after being made up (July 20, 1883), it
developed a ‘little ammonia, which it still contains, but no nitrite or
nitrate. Most ammonia-forming substances, when dissolved in dis-
tilled water, give rise to a little ammonia, unless the solutions are
sterilised by heat and protected from the entrance of ordinary air.
The ammonia-forming ferment, however, does not cause nitrification
even in presence of supply of base.

V. Nitrification of Gelatin.

The sample of gelatin used contained 19°0 per cent. of water.
250 mgrms. were dissolved in 500 c.c. water and 222 mgrms. each of
crystallised magnesium sulphate, sodium phosphate, and potassium
bitartrate, added to the solution. ‘Carbonate of lime and 2 grams
moist soil were also added. The solution belonged to ‘the first series
and was treated in all.respects like its companions.

Mar. 9, 1883. Made up. Total nitrogen present = 36 mgrms.

Mar. 15, 1883. No ammonia.

Mar. 26,1883. Ammonia = 4 mgrms. nitrogen. The solution
turbid, tufts-of mycelium on surface.

April 1,1883. Ammonia = 12 mgrms. nitrogen. Turbidity and
flocks of mycelium much increased, odour putrid.

April 13, 1883. Ammonia = 29 mgrms. nitrogen. No odour;
turbidity lessened:; shows dead bacteria under
microscope. No nitrite or nitrate.

June 4, 1883. NH; present. No nitrite or nitrate.

July 9, 1883. NH; absent. The solution decolorises indigo.
Nitrite present.

July 23, 1883. No ammonia or nitrate. 32 mgrms. nitrogen
present as nitrate.

642 MUNRO: THE FORMATION AND DESTRUCTION OF

This experiment exhibits the ordinary course of nitrification of nitro-
genous organic matter under favourable circumstances. First came a
putrid bacterial fermentation, during which nearly all the nitrogen was
transformed into ammonia. This was succeeded by a pause until the
weather became warm enough for nitrification, which set in some time
between June 4 and July 9, and progressed very rapidly. Nitrite was
at first found, but speedily disappeared, and on July 23, 88 per cent.
of the nitrogen of the gelatine existed as nitrate.

A plain solution of gelatin was also made up March 9, 1883, in
boiling distilled water in a long-necked stoppered flask. The stopper
was removed occasionally when tests were made, so that slight con-
tamination from the air was quite possible. Ina few days the solu-
tion began to grow turbid, developed a putrid odour, and gave a green
colour with Nessler solution. After a little while, the ordinary
ammonia colour was given with Nessler reagent, but the production
of ammonia soon ceased. The turbidity decreased, and the solution
remained without further change until August 21, 1884, on which
date it was nearly clear, but still contained ammonia and retained the
putrid odour. No nitrite or nitrate was present, and in order to see
whether the addition of a salifiable base would produce nitrification a
littie magnesium carbonate was put in. The solution still (March 13,
1886) contains ammonia, but no nitrite or nitrate, thus supporting
the soil-ferment theory of nitrification.

VI. Nitrification of Urine.

Four experiments with diluted urine, I record here because they
present two instances of nitrification through accidental fertilisa-
tion of the solutions. The peculiar behaviour of these solutions, far
from being opposed to the ferment theory of nitrification, is quite in
harmony with it.

El and E2 were made up June 11, 1883. 500 c.c. ammonia-free
water with 2 c.c. urine were placed in two litre flasks, which were
plunged in boiling water for 30 minutes ; the necks of the flasks were
plugged with cotton-wool while steam was issuing. When cold, a
little unsterilised calcium carbonate, in fine powder, was added to
each. El was seeded with soil. E2 was not seeded. Both flasks
were covered with paper caps, and kept in the dark.

In 28 days El contained 2U-——24 mgrms. N as ammonia, and nitrifi-
cation had commenced.

In 50 days 17 mgrms. N were present as nitrite, and oxidation to
nitrate was completed during the next month.

E2, on the contrary, although it developed ammonia nearly as fast as
El, showed no sign of nitrification in 50 days. But after 113 days,

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 643

20 mgrms. N were present as nitrite, which took more than two months
longer to pass into nitrate. There are three possible sources of con-
tamination in this experiment; the unsterilised calcium carbonate,
the entrance of air and dust during the 7 or 8 removals of the
cotton-wool plug for testing purposes, and the use of pipettes
for testing which were washed simply with distilled water instead
of boiling water. The next pair of experiments fixes the respon-
sibility upon the last circumstance, since the two former possible
sources of contamination were present without causing nitrification,
but when the third source was introduced, nitrification followed.

E3 and E4, made up November 10, 1883. Each consisted of 1 c.c.
urine in 100 c.c. water, with a little calcium carbonate. The liquids
were not sterilised by heat, and were placed in stoppered bottles, half
full. E3 was seeded with five drops of the accidentally contaminated
solution E2. E4 was not seeded.

E3 showed signs of nitrification in 37 days, and the nitrite present
gradually increased until on the 155th day it amounted to 12 mgrms. N,
which passed into nitrate before the next testing made on the 225th
day.

The stopper of E4 was not removed during the first 155 days, at the
end of which time neither nitrite nor nitrate could be detected. The
stopper was replaced and not removed again until the 225th day, when
there was still no sign of nitrification. At the third removal of the
stopper, however, on the 283rd day, 8 mgrms. N were found as nitrite,
which had not increased at the fourth testing made on the 368th day.
Up to this time, therefore, the very slight accidental contamination
had caused a very slow, incomplete, and purely nitrous fermentation.
The addition of a centigram of soil on this date, caused rapid and
complete conversion into nitrate.

VII. Nitrification of Ammonium Salts in Artificial Solutions.

The first solutions made up included strong and weak solutions of
ammonium chloride, an intermediate solution of ammonium chloride
not fertilised by soil nor sterilised by heat, and a solution of ammuv-
nium oxalate.

‘Strong and Weak Solutions of Ammonium Chloride.

The weak solution (Al) contained 400 mgrms. of the salt per
litre; the strong solution (A2) 2000 mgrms. per litre. Mr. Waring-
ton’s weak solutions contain generally 80 mgrms. NH,CI per litre,
and when his second report was published (1879), the strongest
solution nitrified contained 640 mgrms. per litre. He states in his
last report (December, 1884) that he has since completely nitrified

644 MUNRO: THE FORMATION AND DESTRUCTION OF

solutions containing 1280 mgrms. per litre, and that solutions con-
taining as much as 8000 mgrms. are in process of nitrification. My
solutions were made up on March 9, 1883; to 500 c.c. of each were
added 222 mgrms. potassiam bitartrate; 222 mgrms. MgSQ, cryst. ;
222 mgrms. Na,HSQ, cryst.; a teaspoonful of calcium carbonate ;
and 2 grams of moist soil. The progress of the solutions (kept in
stoppered bottles, half-full, in a cupboard at the ordinary temperature
of the laboratory) is shown in the following table :-—

March 9, 1883. Madeup. Al contained 52°3 mgrms. N, and A2
contained 262 mgrms. N. Tested and Nesslerised at intervals, neither
of them showed diminution in ammonia, or presence of nitrite or
nitrate up to June 4, 1883. Between this date and July 9, 1883, both
began to nitrify. ‘The subsequent course of the weaker solution Al
was as follows :—

July 9, 1883. Ammonia = 33 mgrms. N. Decolorises indigo.

», 23, 1883. No nitrite. Nitrate = 23 mgrms. N.
Aug. 16,1883. Ammonia = 15 mgrms. N. Nitrite = 9 mgrms. N.
Nitrate = 28 mgrms. N. (by diff.).

Oct. 18, 1883. No ammonia. No nitrite. Nitrate = 47} mgrms. N.

Nitrite was doubtless present between June 4 and July 23, but it
was not tested for; at any rate it had been entirely converted into
nitrate by the latter date, between which and August 16 a fresh
formation of nitrite took place, and the solution was completely
nitrified into nitrate during the long vacation.

The progress of the strong solution (A2), only a third of whose
nitrogen was oxidised to nitrite when Al had run its complete
course, is shown by the estimations of nitrite made at the following

dates :—

July 9, 1883. Nitrite present.
23, 1883. Nitrite= 70 mgrms. N.
. 16, 1883. = = N. Ammonia estimated at
165 mgrms. N.
. 18,1883. - N. Ammonia estimated at
N

”

144 mgrms. N.
, 1884. Nitrite

7 ‘
27, 1884. ) N.
7, 1884. N. Nitrite and nitrate esti-
mated at 115mgrms. N.

Apr.

June 24, 1884, :

Aug. 19, 1884. : é . No ammonia.

Nov. 11, 1884. ; ‘

Mar. 18, 1886. = 86 . Nitrite and nitrate esti-
mated at 236 mgrms. N.

June 5, 1886.

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 645

The completion of nitrification in this solution has only taken place
since I commenced preparing this paper for publication, and alto-
gether the process has occupied more than two years and nine months.
But the solution and residue remaining in the bottle now contain so
much of the nitrifying ferment that 250 c.c. of strong ammonium
chloride solution (1 gram per litre), added bodily, has commenced to
nitrify in five days.

Unfertilised Solutions of Ammonium Chloride, sterilised by Heat,
and not sterilised.

A solution (A3b—March 9, 1883) of ammonium chloride like Al,
but with the soil omitted, underwent a tardy and for a time entirely
nitrous fermentation ; a second solution (A4b) made up November 10,
1883, also unseeded, nitrified completely to nitrite after some time, and
the nitrite has persisted ever since. Both of these solutions were
unsterilised by heat and were kept in ordinary stoppered bottles, half-
full. The nitrification of A4b was much slower than that of a
similar solution (A5) made up at the same time, and seeded with five
drops only of a nitrifying solution (A2), although both nitrified
to nitrite only. I have succeeded in ascertaining definitely that the
standard ammonium chloride* employed in making up A3b and A4b
was itself contaminated with a trace of the nitrous ferment; it had
been made up two years before, and all the solutions in which it was
used which were not sterilised, eventually nitrified. On the other
hand, two boiled solutions (ABl1, AB2) in flasks with cotton-wool
plugs and paper caps, which were made up as checks on April 14,
1884, and not opened until March 18, 1886, have remained without
the formation of any nitrate or nitrite; they still contain plenty of
ammonia.

Nitrification of Ammonium Ozalate.

A solution of the first series (A9, March 9, 1883) contained
250 mgrms. of the crystallised salt. It nitrified to nitrate even more
rapidly than the corresponding solution of ammonium chloride made
up on the same day. I have since ascertained (p. 648) that the oxalic
acid is destroyed long before nitrification sets in. By taking advan-
tage of this fact, ammonium oxalate can be completely nitrified with
half the proportion of salifiable base necessary when ammonium
chloride is used, and sodium or potassium oxalate may be used in
place of the salifiable base usually employed.

* See note, p. 654.
VOL. XLIX.

646 MUNRO: THE FORMATION AND DESTRUCTION OF

Nitrification of Ammonium Chloride; Organic Carbon added in the
form of Potassium Oxalate only.

On July 20, 1883, nine solutions were made up without an alkaline
tartrate; in eight of these, potassium oxalate was substituted; the
ninth, in which ethylamine was to be nitrified instead of ammonia,
contained neither tartrate nor oxalate. The history of five of the
solutions, containing respectively ethylamine, ammonium thiocyanate,
thiocarbamide, urea, and urea (not seeded), has already been given.
The remaining four were devoted to ammonium chloride.

Each of these four solutions contained in 100 c.c. ammonia-free

distilled water—

63 mgrms. ammonium chloride = 16°5 mgrms. N.
20 » MgSO, cryst.

20 » Na,HSO, cryst.

263 ,, potassium oxalate (K.,C,0,).

Three of them, All, Al2b, and Al3c, were rendered alkaline by
magnesium carbonate, and

A11 was fertilised with 1 or 2 mgrms. soil ;

A126 was not fertilised ;

A13c¢ received 1 or 2 mgrms. soil and a few drops of chloroform.

The remaining solution,

Al0, was rendered alkaline by 17-7 mgrms. KHO instead of the
magnesium carbonate; it was fertilised with 1 or 2 mgrms. soil.

The four solutions were kept in 6-0z. stoppered reagent bottles in a
cupboard in a warm room, and were examined at frequent intervals.

Three and a half months after making up, viz., on November 10,
1883, as no nitrite or nitrate had formed in any of them, half of each
solution was removed to a wide-mouthed bottle, a crystal of Rochelle
salt dissolved in each of these halves, and the bottles covered with
loose paper caps. These tartrated halves of the solutions were
examined at intervals up to April 7, 1884, when evaporation and
frequent testings had reduced them to a very small bulk; they were
then thrown away. On that date, nearly eight months after being
made up, and five months after addition of tartrate, neither of them
had commenced to nitrify. AJI3c, from which the chloroform had of
course disappeared long before, had grown a patch of mould, but the
other three were quite free from mould.

The untartrated halves of the solutions, in the stoppered bottles,
were also free from moulds and from nitrite or nitrate on April 7, 1884,
and they were not again examined until August 21, 1884, 13 months
after being made up. On this date, one only of the four had under-
gone partial nitrification ; this was All, which had always been con-

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 647

sidered the most favourable solution, since it contained magnesium
carbonate instead of caustic potash, and was seeded with soil. The
MgCoO, deposit had all disappeared, and the stopper jumped out when
loosened, from accumulation of carbon dioxide gas. Ammonia was
still present, but the reaction of the liquid, originally alkaline, was
now faintly acid. 10 mgrms. nitrogen were present as nitrite, so that
nitrification had evidently proceeded until all the available base
was saturated. Some more MgCO,; was added at this date, and
the solution again examined on November 13, 1884, when all the
ammonia had disappeared, and 144 mgrms. nitrogen were present as
nitrite, out of the 16} mgrms. originally added as ammonia. The
nitrite has persisted without alteration for another 16 months, so that
the fermentation in this solution was purely nitrous throughout.

Al10, the caustic potash solution, was re-seeded with a few milli-
grams of soil on August 14, 1884, and by November 13, 1884, this also
had nitrified entirely to nitrite, and the nitrite still persists without
diminution (March 19, 1886).

To A12b, the unseeded solution, a little fresh MgCO; was added on
August 21, 1884, but up to the present date no nitrification whatever
has taken place, and the solution gives a deep brown precipitate with
Nessler’s test. Here we have another instance of a non-fertilised
but also non-sterilised solution, refusing to nitrify for two years and
four months, although exposed to occasional chances of contamination
from the air.

To Al3e, the fertilised but chloroformed solution, some more soil
was added August 21, 1884. On November 13, 1884, the odour of
chloroform was very faint, but no nitrification had occurred. To-day
(March 19, 1886) there is no odour of chloroform, but still there is no
nitrification, and the solution gives an abundant brown precipitate
with Nessler test. Of these four solutions then—

All, with MgCO; as base, underwent a complete nitrous fermenta-
tion without re-seeding, but after a long delay.

A10, with KHO as base, underwent a complete nitrous fermentation,
but only after re-seeding.

A12b, with MgCO, as base, but not fertilised, has altogether refused
to nitrify.

Al3c, with MgCO, as base, fertilised and chloroformed, has alto-
gether refused to nitrify, even after re-seeding.

These solutions, therefore, strongly support the ferment theory of
nitrification.

As regards the substitution of an oxalate for a tartrate, it remains
for me to add that in four out of the eight solutions to which it was
added, the oxalate had comp'tetely disappeared one month afterwards

2x2

648 MUNRO: THE FORMATION AND DESTRUCTION OF

(August 16, 1883). Three out of these four solutions had been seeded
with soil. The two solutions in which the oxalate persisted unaltered
for some months, were the wnfertilised solution A12b and the chloro-
formed solution Al3c.

These facts suggested to me that the oxalate was decomposed by
the agency of soil ferments, and hence that its “ organic carbon ”
exercised no influence on the nitrification. I have recently verified
this. Solutions of alkaline oxalates are rapidly attacked by soil
ferments; they become opalescent with bacteria, the oxalate is
destroyed in a few days, and an alkaline carbonate is found in its
place. This change is entirely prevented by boiling, and to a great
extent by chloroform. It follows from this that an alkaline oxalate
may be used as salifiable base in nitrification experiments. For
example, 100 mgrms. ammonium oxalate crystals were completely
nitrified with 170 mgrms. sodium oxalate as salifiable base. The
oxalate was entirely destroyed by the 5th day ; nitrification commenced
between the 21st and 26th day; three-fourths of the nitrogen was
present as nitrite on the 65th day, and nitrification to nitrate was
finished before the 80th day.

Nitrification in boiled, filtered, transparent solutions of Ammonium
Chloride.

Three solutions were made up on November 10, 1883, with the
especial object of studying the influence of nitrification on the trans-
parency of a solution and of observing the formation or non-forma-
tion of a deposit or surface growth. Sodium bicarbonate was, there-
fore, used as the salifiable base instead of calcium or magnesium
carbonates; the solution was filtered twice through Swedish paper,
boiled, and protected from the entrance of unfiltered air; and ferti-
lisation was effected in one case by only 1 or 2 mgrms. soil, and in the
other case by five drops of an apparently clear solution undergoing the
nitrous fermentation (A2). The third solution was not fertilised and
was not opened for six months after making up. Each solution
contained in 100 c.c. ammonia-free distilled water—

63 mgrms. NH,Cl = 16°5 mgrms. N.
40 » Rochelle salt cryst.

10 » MgSO, cryst.

10 » Na,HPO, cryst.

60 » NaHCO.

These solutions were filtered through double Swedish filters into

narrow-necked 200 c.c. flasks, previously cleansed with boiling pure
sulphuric acid. The solutions were sterilised by boiling, the necks of

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 649

the flasks were plugged with cotton-wool while steam was issuing,
and, on cooling, paper caps were tied over the mouths. Of course
some ammonia was lost during the boiling, and also subsequently by
evaporation through the cotton-wool plugs, but plenty remained
intact and underwent nitrification in due course.

The flasks were locked in a bureau in a living room, and were
examined as to transparency, &c., from time to time. The cotton-
wool plugs of two of them, A6 and A7, were removed on the cooling
of the solutions for the introduction of the ferment, and at intervals
afterwards in order to mark by testing the commencement of nitri-
fication ; the plug and cap of A8b, the unfertilised solution, remained
untouched from November 10, 1883, to April 14, 1884. The following
observations were made at the dates specified :—

Nov. 20, 1883. The solutions A6 (fertilised with a mgrm. of soil)
and A7 (fertilised with five drops of a nitrifying solution) are both
opalescent. There is no sign of nitrification. The opalescence of A6
is seen under the microscope (,'," objective) to be caused by bacteria,
which exhibit a slow oscillating motion, a few rotifers are also present
in active motion.

Dec. 3, 1883. A6 has become clearer; A7 still turbid.

Dec. 9, 1883. A6 still clearer; dead bacteria are settling down in
minute patches. A7 still turbid. No nitrifica-
tion in either.

Jan. 24, 1884. A6 clear, with deposit of dead bacteria. Nitrifica-
tion has commenced.

A7 nearly clear, with deposit of dead bacteria ;
three or four tufts of mould have commenced to
grow from the bottom. Nitrification has com-
menced.

Feb. 15, 1884. A6, appearance unaltered. Contains nitrite.

A7 not quite clear. Clouds of mould increased.
Contains nitrite.

April 9, 1884. A6 still contains ammonia. Nitrogen as nitrite =
1} mgrm. Nitrate also present, but no exact
estimation could be made.

A7 still containsammonia. Nitrogen as nitrite =
2 mgrms. Doubtful whether any nitrate pre-
sent.

Up to this date, therefore, A6 had undergone a partial fermenta-
tion into nitrite and nitrate, and A7 a partial fermentation into
nitrite. The unfertilised solution A8b, opened for the first time
April 14, 1884, was absolutely bright and clear, and contained not a
trace of nitrite or nitrate. Compared with this solution, the clearness

a a

650 MUNRO: THE FORMATION AND DESTRUCTION OF
of A6 and A7 was seen to be far from perfect ; A6 was the clearer o.
the two, but was faintly opalescent.

June 24, 1884. A6 and A7 still contained ammonia; nitrite dimi-
nished to 4 to 1} mgrm. N respectively.

The unfertilised solution A8b did not remain absolutely bright
after being opened on April 9. It soon developed a delicate
opalescence, which remained on June 24, 1884. There was then no
sign of nitrification, however, so that the first temporary exposure to
the air, although it had certainly introduced germs into the solution,
had not inoculated it with the nitrifying ferment.

Aug. 21, 1884. By this time A8b had begun to nitrify, so that the
second exposure or testing on June 24 was fatal toit. Ammonia
was still present. The N as nitrite was 2°8 mgrms., and a very care-
ful estimation (after oxidation with permanganate) of the total N as
nitrate gave 2°77 mgrms. Evidently, therefore, the fermentation was
purely nitrous. The opalescence previously noticed had ended in a
slight flocculent deposit.

A6. The nitrite has now disappeared, but ammonia is still present.

A7. The nitrite has increased to 5 mgrms. N, but ammonia is still
present.

Nitrification in A6 and A7 being arrested by saturation of the
sodium bicarbonate, more of this salt was added to each solution on
this date. By November 13, 1884, A6 had nitrified completely to nitrate ;
A7 afterwards nitrified completely to nitrite; in A8b, the formation
of nitrite has proceeded more slowly, and even now a little ammonia
is unnitrified.

Of these solutions, therefore, the only one that has, in 32 months,
completely nitrified to nitrate is the one seeded with soil ; the solution
seeded with a nitrous fermenting solution, has nitritied completely to
nitrite; and the unseeded solution, which remained absolutely
intact so long as it was not exposed to contamination, has partially
nitrified to nitrite. These results are in exact accordance with Mr.
Warington’s conclusions, and are scarcely explicable by any other
theory than that of ferment introduced from without.

The difficulty in producing nitrification in otherwise clear tartrated
solations will be seen from the above description to be very great.
The organisms which render so many organic infusions opalescent
after even a momentary exposure to unfiltered air, are immensely
more abundant than the nitric ferment, and very much easier of
cultivation; and they not only attack the organic matter and
render the solutions turbid, but they reduce and sometimes destroy
any nitrate that may be present in the liquid. Unless ordinary
bacterial germs are rigorously excluded, therefore, the presence
of organic matter of this sort must be prejudicial to nitrification.

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 651

This exclusion may possibly be accomplished by employing a very
pure cultivation of the nitrifying organism, and by permitting
the access of filtered air only. Buta much easier method of avoid-
ing the multiplication of these bacteria is to exclude Rochelle salt or
any similar organic matter from the solutions ; and this may be done
either by relying on the organic carbon in the soil used for seeding,
or by using, as a nitrifying medium, well water to which ammonia
has been added. The organic matter of soil and of well water, being the
residue of bacterial fermentations, is impervious to the attack of the
bacteria in question, and does not encourage their multiplication. Hence
such solutions remain clear, although seeded with soil and exposed to
unfiltered air.

VIII. Is Organic Carbon essential to Nitrification ?

This is a vital question, for could it be positively shown that nitrifi-
cation will take place in the absolute absence of organic carbon,
either the theory of an organised nitrifying ferment would fall to the
ground or we should have to admit two processes of nitrification, one
depending on, and the other independent of, a nitrifying organism.
There is no point, however, in the whole study which is surrounded
with more difficulties than this one. For granting that we succeed in
excluding accidental organic matter from the distilled water, the
ammonium chloride, the salifiable base, and the other minerals em-
ployed, we have still to deal with the air, perhaps filtered through a
cotton-wool plug, which in itself may be a source of contamination.
Even supposing these difficulties overcome, there remains the organic
matter introduced with the seed, which may be either a few particles
of soil ora few drops of a nitrifying solution containing added
organic matter.

I therefore discuss the subject more with the view of throwing
light on the practical requirements of the process of nitrification as
regards organic carbon than with any intention of making the pre-
sence or absence of carbon a crux of the organism theory of nitrifica-
tion. This theory is susceptible of other and easier lines of proof,
and to my mind is supported by overwhelming evidence.

In nearly all of Mr. Warington’s published experiments, either an
alkaline tartrate (in a few cases glucose) has been added to the solu-
tions with a view of supplying carbonaceous aliment for the nitrifying
organism, or else a substance has been chosen for nitrification which
in itself contained an abundant supply of organic carbon (asparagine,
milk, rape cake, urine, &c.). There are, however, two experiments,
recorded in his 1879 paper, in which nitrification occurred in the pre-
sence of very little carbon. These solutions contained ammonium

652 MUNRO: THE FORMATION AND DESTRUCTION OF

chloride, potassium sulphate, calcium phosphate and carbonate; some
organic carbon was, however, added, for they were seeded with a
small fragment of mushroom mycelium, and with 1 c.c. of a shaken
up liquid which had been seeded with 14 grams of soil and had re-
ceived 50 mgrms. potassium bitartrate as well. One of these (kept in
the dark) nitrified completely to nitrate in nine months; the other, in
the light, nitrified partially to nitrite. These I believe are the only
recorded experiments in which the proportion of organic carbon to
ammoniacal nitrogen was less than 3 : 10, exclusive of the carbon in
the seed, and in most of them it was much greater. In his third
(1884) report, Mr. Warington says that “ probably tartrate 100 per
cent. of the ammonium salt is sufficient for every purpose,” and that
“urine apparently contains quite enough organic carbon for its own
nitrification.”

Amongst the experiments already described will be foand—

One (B3) in which 31 mgrms. ethylamine hydrochloride were par-
tially nitrified into nitrate after twice seeding with a few milligrams
soil ;

One (G1) in which 35 mgrms. ammonium thiocyanate with
20 mgrms. oxalic acid crystals were completely nitrified into nitrate
after twice seeding with a few milligrams soil ;

One (Fl) in which 20 mgrms. urea with 20 mgrms. cryst. oxalic
acid were completely nitrified to nitrate after twice seeding with a few
milligrams soil ;

One (All) in which 63 mgrms. ammonium chloride with 263
mgrms. potassium oxalate were completely nitrified to nitrite after
once seeding with 1 or 2 mgrms. soil ;

One (Al10) in which 63 mgrms. ammonium chloride with 26°3
mgrms. potassium oxalate were completely nitrified to nitrite after
twice seeding with 1 or 2 mgrms. soil.

I now know that the oxalate in these experiments had nothing to
do with the nitrification, since it was destroyed long before nitrifica-
tion set in.

The experiments with natural waters, presently to be described,
will show that the organic carbon in well water is sutlicient for the
nitrification of added ammonia in quantity at least equal to the
saturating power of the calcium carbonate present for the nitrous or
nitric acid produced. The series of experiments which now follows,
however, is more to the point, and pretty clearly exhibits the practical
relation of organic carbon to nitrification.

Six solutions were made up on February 9 and 14, 1884. All con-
tained ammonium chloride equal to 40 mgrms. NH; per litre
(33 mgrms. NN). They were carefully filtered after being made up,
but were not sterilised by heat. Al4b measured 500 c.c., and was con-

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 653

tained in a litre flask, the neck of which was plugged with cotton-
wool; the others measured 100 c.c. each, and were contained in 6-oz.
stoppered reagent bottles. The following table shows the arrange-
ment adopted :—

Feb. 9, 1884. A18. Complete solution, containing per litre 126

mgrms. NH,Cl, 200 mgrms. NaHCO, minerals
(40 mgrms. MgSQ, cryst., and 40 mgrms. K;PQ,)
40 mgrms. Rochelle salt; seeded with a few
milligrams soil.

Al7. Rochelle salt only omitted.

Al6b. Soil only omitted.

Al9. Rochelle salt only omitted ; seeded with several
grams of soil.

“ - Al15b. Rochelle salt and soil omitted.

Feb. 4, 1884 A146. Rochelle salt, soil, minerals, and salifiable base

omitted.

Five of these solutions, including the two unseeded ones, deve-
loped nitrite or nitrate sooner or later; but they exhibited marked
differences in rapidity and extent of nitrification.

A19 nitritied completely to nitrate in less than 50 days.

A17 also rapidly nitrified; half the nitrogen was present as nitrate
in 30 days; a little nitrite was formed and afterwards disappeared.

Al18, differing from the last only in containing Rochelle salt,
nitrified more slowly. In 30 days, less NH; had nitrified, and the
nitrite present was considerable; during the summer, it passed com-
pletely into nitrate.

A16d, with Rochelle salt but no soil, underwent at first a slow and
entirely nitrous fermentation ; 14 mgrms. of nitrogen were present as
nitrite even on November 13, 1884, but some time after that date con-
version into nitrate took place.

Al5b, with no tartrate or soil, and therefore without added organic
matter of any sort, nitrified like the last, but still less readily.
12} mgrms. nitrogen were present as nitrite on August 17, 1884; this
increased to 15} mgrms. on November 11, 1884, and has remained
without further oxidation ever since. The clearness of the solution
has never been visibly impaired.

Al4b, without minerals, organic matter, or salifiable base, was
tested five times at intervals from February 4, 1884, to August 21,
1884, but developed no nitrite or nitrate. A little sodium bicarbonate
was then added, and a trace of nitrite was found on November l,
1884, which did not increase. Seventeen months later, although the
trace of nitrite was still present, 10 c.c. of the solution would not de-
colorise three drops of standard permanganate or three drops of indi-
gotin (1 c.c. = 0°035 mgrm. N).

654 MUNRO: THE FORMATION AND DESTRUCTION OF

As regards the source of contamination in the three unseeded but
unsterilised solutions, Al4b, Al5b, Al6b, I have lately succeeded in
proving decisively that the standard solution of ammonium chloride
used, which was not boiled, actually contained the nitrous ferment.*
The source of inoculation being known, it is not a matter of surprise
that fermentation in these solutions should have taken place, but the
differences in their behaviour are very significant. Al4b with no
added organic matter or minerals, develops only a trace of nitrite,
although kept for 17 months after addition of a salifiable base. On
the other hand A158, with minerals but no added organic matter, under-
goes a complete but purely nitrous fermentation ; and A16b, for the first
nine months at any rate, follows precisely the same course, although
tartrate was purposely added. The merest traces of organic matter,
therefore, such as may be furnished accidentally by occasional expo-
sure to the dust of the air, are sufficient for complete nitrous fermen-
tation of the quantities of ammonia used in these experiments ; but it
does not appear that the process is equally independent of purposely
added minerals.

A17, again, shows that the organic matter of 1 or 2 mgrms. of soil
suffices for the requirement of a complete nitric fermentation; and
the addition of tartrate to Al8 was not attended with the smallest
advantage.

Moreover, unless aérial bacteria can be rigidly excluded, and nitrifi-
cation carried on with pure cultivations of the nitric ferment instead of
soil, the presence of a tartrate or any similar organic body is distinctly
prejudicial to nitrification. This will appear clearly from the experi-
ments detailed in the next section.

IX. Nitrification of Ammonium Salts in Natural Waters.

It is a matter of common knowledge that the ammonia and nitro-
genous organic matter which find their way, say in the form of sewage,
into our rivers and wells, ultimately give rise to the formation of
nitrates; and the proportion of “free ammonia” and “ nitrogen as
nitrite and nitrate” have had their place in our analytical schemes
as measures of sewage pollution “ present” and “ past.” But few
investigations, however, have taken place as to the exact manner in
which the transformation of ammonia into nitrite or nitrate is
effected; and such as have been made have proceeded upon purely
chemical lines, the rapid aération and large surface exposure which

* This standard solution, made up more than four years ago, now contains a trace
of nitrite sufficient to give an orange colour with metaphenylenediamine; and by
adding a little of it to sterilised well water, I have recently been able to effect com-
plete conversion of the ammonia into nitrite.

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 655

moving water undergoes being tacitly or openly inferred to be sufii-
cient causes of nitrification.

As soon as I had repeated some of Mr. Warington’s experiments
which strongly supported the theory of a special nitrifying organism
in the soil, 1 commenced (November, 1883) a series of experiments
with the view of ascertaining whether natural waters oxidise
ammonia because of their contact with soil and the ferment con-
tained therein, or because they themselves contain this ferment which
they have obtained from the soil, or (according to then current theories)
simply because they offer the ammonia the means of free and thorough
atmospheric contact. The experiments were continued in 1884, but
(until within the last two months) no new solutions were started
after April of that year, although those then in operation have been
under observation ever since in order to ascertain the completion or
non-completion of nitrification in some of them.

Some months after my first experiments on waters were com-
menced, the late Dr. Angus Smith’s last Report to the Local
Government Board was published, and although he mentions some
facts as to denitrification which he had observed some time before,
and with which I had then become familiar from my own experi-
ments, he does not appear to regard the oxidation of ammonia in
waters in any other light than a purely chemical one. In his previous
Report, indeed (1882), he says : “ Putrefaction and oxidation are two
well-known modes of destroying organic bodies at ordinary tempera-
tures. The second is not proved to be connected with organisms.”
In the last Report he says : “ Animal or vegetable matter containing
nitrogen produces nitrates by oxidation with and without organisms.”
He illustrates the latter position by quoting experiments on the oxida-
tion of ammonia by permanganates, and by persalts of iron, and seems
to regard it as the more important process, for he says that it is “an
action of the oxygen without the intervention of organisms probably
carried out to a great extent in nature.” Whilst fully admitting
that ammonia may be oxidised by the very powerful reagents named,
it is much to be doubted whether a purely chemical oxidation of
ammonia ever takes place in nature when organic matter decays in
contact with air, water, or soil. As regards the ordinary oxidation
of nitrogenous organic matter following putrefaction, Warington’s
experiments with rape cake, asparagine, milk, urine, &c., show that
the presence of the nitrifying ferment of the soil is a sine gud non
of nitrification; and the experiments which follow show the same
statement to be true of the ammonia contained in natural waters.

My experiments were made by adding known quantities of ammo-
nium chloride to waters under various conditions ; and it occurred to
me that if waters possessed the power of readily nitrifying ammonia

656 MUNRO: THE FORMATION AND DESTRUCTION OF

added in this way, they would form very suitable media for the study
of nitrification and for the culture of the nitrifying ferment, the
organic matter and minerals necessary for the nutrition of the
ferment, as well as the salifiable carbonate of lime, being present in
such forms as not to interfere with the clearness of the liquid.
A still greater advantage is that well waters have no tendency to
encourage the growth of the bacteria which multiply in and render
turbid most organic infusions during the first week or two of their
exposure to the air. The germs of these bacteria are present in all
well waters, but they do not multiply because of the absence of
suitable organic matter. There is one drawback to the use of well
water for experiments on nitrification. In most cases it already con-
tains nitrate. In my experiments the proportion of nitrogen as nitrate
was therefore very carefully determined beforehand, and the progress
of the nitrification of the added ammonium salt was ascertained by—

Testing at intervals with Nessler reagent for the disappearance of
the added ammonia ;

Testing with metaphenylenediamine for the appearance of nitrite,
and estimating the quantity in some cases by standard potassium
permanganate ;

Determining by means of standard indigotin (Warington’s im-
proved process) any increase over the original nitrate present,—
in the presence of nitrite and nitrate, the nitrite was first esti-
mated with permanganate, and the total nitrogen then present as
nitrate estimated with indigo.

The merest traces (if any) of free ammonia and nitrite were present
in the waters with which I worked; and even the organic nitrogen
present was in altogether insignificant proportion to the ammonia
salt added, and to the ammonia actually nitrified; it did not exceed
zioth part of the added ammonia.

The standard solutions employed were of the following strength :—

Potassium permanganate... 1 c.c. = 0°072 mgrm. N as nitrite.
Indigotin (03847 - nitrate.
= 0°055 ” ”
(Diluted to 4, 4, &c., according to circumstances.)

Ammonium chloride, 1 c.c. = | mgrm. NH;.

Experiments with River Water.

The earliest experiments showed the rapidity and ease with which
relatively large quantities of ammonium chloride could be nitrified in
natural waters. Three dark glass bottles were half filled with 500 c.c.

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 657

each of water from the River Avon, which irrigates the water meadows
of the whole valley, and is frequently analysed in the College of
Agriculture Laboratory.*

To each bottle was added 25 c.c. standard ammonium chloride
= 20°6 mgrm. N.

Received no addition beyond the ammonium chloride.
A crystal of Rochelle salt.
- ™ and a centigram of soil.

When made up on November 21, 1883, 10 c.c. of each of these
solutions decolorised 0°53 c.c. indigotin (1 c.c. = 0°055 mgrm. N), and
contained, therefore 0°204 grain N per gallon as nitrate.t The
bottles were simply corked, and were kept in the dark in a bureau by
the side of a fire-place in a living room.

8 and y clouded in a few days; a remained clear.

Tested on December 17, 1883, 26 days after making up, 10 c.c. a
decolorised 7°4 c.c. indigotin. A little ammonia was present, but no
nitrite.

Nitrogen originally present as nitrate = 0°204 grain per gallon.

- in added ammonia in -

_——

2°951

- as nitrate on Dec. 17, 1883 = 2°849

2°6 grains of nitrogen in NH,Cl require for nitrification 18°5 grains
per gallon of calcium carbonate, whereas the water contained only
15 grains ; doubtless the magnesia, potash, and traces of other bases
in the water supplied a grain or two more of salifiable base.

* Analyses made at various times show the composition of this water to be as
under :—
Total solid matter varies from 19°98 to 21°32 grains per gallon.
Free ammonia ‘i O ,, trace.
* Albuminoid ”
varies from....
Nitrogen as nitrate variesfrom 0°09 ,, 0°46 grain per gallon.
- nitrite O ,, trace.
CO, 6°4 123 grains
Calcium carbonate 14°58 ,, 15°26
Loss on ignition 2°50 3°85
Chlorine 0°92 1:05
Calcium sulphate O ,, trace.

0°08 part per million.

+ It will be understood that the quantities of water taken for analysis, and the
dilution of the indigotin, were varied as is necessary in working the process ; but for
the sake of uniformity all the results are calculated for 10 c.c. water and for indi-
gotin, of which 1 c.c. = 0°055 mgrm. nitrogen, the strongest solution used.

658 MUNRO: THE FORMATION AND DESTRUCTION OF

Not only had £ and y not nitrified in the same time, but y (soil and
Rochelle salt) had lost every trace of the nitrate it originally con-
tained, and £ had lost the greater portion.

Both of these solutions eventually nitrified, but the addition of the
Rochelle salt, which determined denitrification, brings them more
appropriately under that head, to be considered by-and-bye. At
this stage they showed me that I could with advantage eliminate
Rochelle salt from my experiments, and rely on the organic carbon of
the water only.

Experiments with Laboratory Well Water, commenced December 19,
1883.

A series of experiments was next made with the water of the Labo-
ratory Well, an unpolluted but very shallow well sunk in alluvial
gravel, and situated about 400 yards from the River Avon.

This well water contains—

Total solid matter ........ 23'1 grains per gallon.
PD sadosenennceeece 1:0 ‘ -
Free ammonia ............ none

Nitrogen as nitrate........ 04144 ,, -

» BED cccccecs none.
ere 6—13 ,, ”
Sulphates........esceceee none.

Calcium carbonate. ........ 54 ~=,, ~

2500 ¢c.c. of this water were placed in a half-gallon ‘ Win-
chester quart” and 100 c.c. standard ammonium chloride added
(= 100 mgrms. NH; = 82°4 mgrms. N). The water was titrated with
indigotin after this addition, when 10 c.c, decolorised 1:0 c.c. indigotin.

Nitrogen as nitrate ......... 0°427 grain per gallon.
Nitrogen added as ammonium
GRISTEED co cccccccvcceces 2°217 grains -
2644, %

This was divided between 16 flasks and bottles, as in the following
table :—

With no added Organic Mutter.
L16. Unsterilised.
L4. Chloroformed.
Ll. Boiled, and not opened or tested until April 14, 1884.
L3 Boiled, and then seeded with a few milligrams soil.

659

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS.

. Boiled, then seeded with five drops of nitrifying solution (A2).
L2. Boiled, then seeded with five drops unboiled well water.

With added Organic Matter.

L7. Rochelle salt only added.
L8. Cane-sugar only added.
L9. Glycerol only added.
L10. Sodium acetate only added.
L5. Rochelle salt added ; boiled, and not opened or tested until
April 14, 1884.
L6. Rochelle salt added ; boiled, then seeded with soil.
L1l. Rochelle salt and a little phenol added. Seeded with soil.
L13. Salicylic acid added ; seeded with soil.
L12. Rochelle salt and a little chloroform added. Seeded with soil.

L 1, 2, 3, 5, 6, 15, the boiled solutions, were kept in flasks, half full,
the necks of which were plugged with cotton-wool, and a paper cap
tied over the mouths. The seed, if any, was added during a
momentary withdrawal of the plug, after cooling.

The remaining solutions were kept in stoppered reagent bottles,
half full, except L16, which formed a layer of about 2 inches depth at
the bottom of a stoppered Winchester quart.

Of these 15 solutions five, viz., the two boiled and unseeded solu-
tious, L1 and L5, the two chloroformed solutions, L4 and L12, and
the solution with phenol, L11, absolutely refused to nitrify, although
tested at intervals for more than two years after being made up.
Filtered air had access to these waters through the cotton-wool
plugs all that time.

Nitrification, to a greater or less extent, took place in three or four
of the solutions containing added organic matter, but since in all
these solutions partial or complete denitrification of the water was
first produced, they are better considered under the section devoted
to denitrification (p. 667).

The remaining four solutions, free from added organic matter, all
nitrified completely to nitrite or nitrate, but at different rates.

L15, seeded with five drops of a strong solution in active nitrifica-
tion, underwent almost complete oxidation to nitrate in less than two
months ; for on February 25, 1884, a trace only of nitrite was present,
and nearly 3 grains per gallon of N existed as nitrate.

L3, seeded with a few milligrams of soil, completely nitrified to
nitrate in about the same time as L15. Nitrite was present as late
as February 4, 1884.

L16, the unboiled well water with ammonium chloride only added,
completely nitrified to nitrite in less than 54 days, but the nitrite bas

660 MUNRO: THE FORMATION AND DESTRUCTION OF

since remained without any further alteration for more than two years.
The following table shows the dates of testing :—

Dec. 19, 1883, 10 c.c. consume 1:1 c.c. indigotin.
Jan. 24, 1884, 10 ‘i 15 a 7
Feb. 4, 10 2°6
9 20, 10 2°6
Mar. 15, 10 2°6
Apr. 17, 10 2°6
Aug. 21, 10 2°6 J Nitrite = 1:96
grains N per gallon.

> Nitrite present.

Nov. 11, 1884, nitrite undiminished.

Aug. 14, 1885, _,, -

Mar. 27, 1886, __,, - A little ammonia still present.
Nitrogen as nitrite and nitrate = 2°11 grains per gallon.

2°22 grains of added nitrogen (as NH,Cl) require 15°8 grains per
gallon CaCO, for complete nitrification, which is rather more than that
contained in the water; hence a little ammonia escaped nitrification.
The persistence of the nitrite for over two years is a very striking
result, and points to an essential difference between the nitrifying
power of this well water and that of the soil and nitrifying solution
used in L3 and L15.

L2, boiled and seeded with 5—10 drops of the unboiled well water,
also nitrified, but very slowly, and, as in the last case, nitrite only
was produced. The formation of nitrite, instead of being complete in
less than 54 days, had then made little progress and was not finished
until two months later. The nitrite (1°96 grain N per gallon) has
persisted for over two years.

Four more experiments were made with this well water, in order to
see whether filtration would remove the nitrous ferment which it
appeared to contain, and to again compare its behaviour with a soil-
seeded water.

Each of the four solutions, L17, 18, 19, 20, consisted of 500 c.c.
laboratory well water + 20 c.c. standard ammonium chloride
(= 20 mgrms. NH; = 63 mgrms. NH,Cl). They were treated as
follows :—

L17 and 18 were filtered through two Swedish filters into cleansed
and sterilised 1500 c.c. flasks, which were plugged with cotton-wool.
But whereas L17 was opened and tested at short intervals, L18 was
not opened for 40 days after being made up.

L19 was similar to the two preceding, but was seeded with a few
milligrams soil.

L20 consisted of the unfiltered mixture, and was placed in a stop-

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 661

pered bottle to allow of accurate determinations of nitrite from time

to time.
The subsequent course of these solutions is shown in the following

table :—

Consumed by 10 c.c. of each wa‘er.

Made up c.c. Indigotin.

1884. Description.

c.c. Permanganate:
August 26, 1884.

Feb. | en, | Bean
4&9| 25. | 20.

7. Filtered ssienenceil 1°4 | 1°7 | 2°9|4°0 = 196 grains N per
| gallon as nitrite.
. Filtered and un-| 14 | — | 2°9 “e = 196 grains N per
opened until | gallon as-nitrite.
March 20 .
. Filtered and seed-| 1°4 | 1°6; 1°8 |3°5 = 1°72 grains N per
ed with soil | | gallon as-nitrite.
. Unfiltered ..... | 1°4 | 1°4 | 3°4 | 3°4 = 1°67 grains N
| | | | gallon as nitrite.
|

|
|

Nitrogen added as ammonium chloride = 2°22 grains per gallon.

L17, 18, 19, being in open flasks, were subject to. evaporation.
Careful estimations were made in L20 at. intervals for two years,
without showing the slightest further change. The: latest determina-
tions give—

Nitrogen as nitrite 1-67 grains per gallon.
nitrite and nitrate.... 1°96 “ - °

nitrate 0°29 by difference.

As there was about 0°54 grain per gallon of nitrogen as nitrate in
the original water, it will be:seen that the fermentation was purely
nitrous.

L17 and L18 followed precisely: thé same course, showing that
filtration did not remove the nitrous ferment, although it appears to
have weakened it; the unfiltered solution had made more progress in
nitrification on March 20 than the two filtered ones. The period of
incubation in these solutions was under 2] days; temperature
G0—70° F.

L19 presents an apparent anomaly. Although seeded with soil, it
nitrified less rapidly than the rest, and to nitrite only. But all these
bottles were kept in diffused daylight, and L19 slowly grew a powdery
green alga, the spores of which were introduced with the soil. By ihe

* Ammonia is still present, because the salifiable base is exhausted.
VOL. XLIX. 2y¥

662 MUNRO: THE FORMATION AND DESTRUCTION OF

continued growth of this alga, all the nitrite originally formed was
gradually destroyed in the course of a year or so.

¢ The experiments hitherto described show that whereas distilled
water, even after the addition of a salifiable base, will not nitrify added
ammonium chloride (A14b), well or river water fertilised with a
particle 6f:soil or a drop of nitrifying solution supplies everything
necessary for the nitrification of added ammonium chloride to the
full extent of' the saturating power of the calcium carbonate present ;
and that nitrification in these waters, as in the artificial solutions used
by Warington, is entirely prevented by boiling or by the addition of
an antiseptic. But they further appear to show that river water itself is
sufficiently contaminated with this soil ferment to produce a complete
nitric fermentation of added ammonia; and that a well water, in five
separate experiments, possessed sufficient nitrifying power to produce
a complete nitrous fermentation. Also, that whilst this nitrifying
power is perhaps weakened by filtration of the water through Swedish
paper, it is not entirely removed.

Since Warington’s experiments (1884 Report) prove that the
nitrifying organism is confined to the surface soil, the question as to
how far different natural waters become contaminated with the
ferment is one that can only be settled by experiment. A priori, we
should suppose that all surface waters, and all waters liable to
accidental contamination with soil, must possess this ferment ; but it
is possible that they do not all contain sufficient for the practical
requirements of nitrification, and even that some protected deep well
waters may be quite free from it. Quite recently, therefore, I have
re-tested the well and river waters already alluded to, and have
examined several other natural waters, under the following rigorous
conditions :—

A flask, holding about 350 c.c., is carefully cleansed, and distilled
water is then boiled in it, and a cotton-wool plug inserted in the neck.
After a few minutes’ boiling the plug is removed for a moment and
the distilled water replaced by 5 c.c. standard ammonium chloride
(= 5 mgrms. NH;). The plug is again inserted, the ammonium
chloride boiled, and the flask covered with a paper cap and allowed
to cool. When cold it is taken to the water supply, the cap and plug
carefully removed, and about 300 c.c. water placed in the flask, either
directly from a pump or tap, or by means of a sterilised beaker. The
plug is then replaced, the paper cap tied over the mouth, and the
flask is placed in a warm place in the dark, and not opened until a
sufficient period has elapsed to allow of the commencement of nitrification.
If nitrification has commenced, nitrite will then in nearly all cases be
present, and will give a yellow, orange, amber, or red colour with
the metaphenylenediamine test, according to the amount. Should

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 663

complete nitrification have taken place, the ammonia will have dis-
appeared; and should a partial but purely nitric fermentation have
occurred, more indigo will be consumed than was the case with the
original water. (In experiments with well waters unseeded with soil,
I have hitherto invariably found nitrite present during the greater
part of the nitrifying period.)

The following table exhibits the results obtained with six river and
‘shallow well waters examined at Downton, and seven miscellaneous
waters examined at Bristol. The samples altogether comprise shallow
and deep well waters, pure waters, and waters with “present” and
with “past” sewage contamination, the water of a “level” in a coal
mining district, river water, and rain water. A summary of the
composition of each water is given, and the result of the meta-
phenylenediamine test at the conclusion of the period allowed for
incubation, and afterwards. The Bristol] samples were kept at a tem-
perature of 80—85° F., excepting the boiled or blank solutiun. This
one belonged to a series made up May 7—20, 1886, but unfortunately
they were slightly over-heated during a portion of the incubating
period ; the temperature rose to a few degrees over 100 F. on two or
three days, and as a consequence all except one refused to nitrify.*
That they were really sterilised by the heat is proved by the fact that
duplicate flasks, made up as soon as the overheating was discovered,
all began nitrifying in less than 18 days. These sterilised flasks,
therefore, strengthen the evidence afforded by the boiled or blank
solution. The Downton waters were incubated at the ordinary tempe-
rature, and with one exception commenced nitrification in less
than 18 days; and this commenced nitrifying between the 18th and
23rd day.

The “rain water caught during a shower in a sterilised beaker ”
was of course added to some boiled well water, since of itself it did
not contain the salifiable base and other impurities essential to
nitrification.

The shorter of the two intervals mentioned under each water is
the date of first opening and testing the flasks; a second testing was
made after another few days in order to ascertain the relative
rapidity of nitrification in the different waters. The colour with
metaphenylenediamine on the first testing is no guide, since the
nitrite increases during the first period of nitrification and decreases
during the second, and itis necessary to know whether the nitrite is in
the increasing or decreasing phase. The second testing accomplishes
this. Thus with the Downton waters we have (see p. 666)—

* This one began to nitrify before the over-heating took place, and the nitrifica-
tion has made no further progress since. A second testing, made 40 days after the
first, still shows no nitrification in any of the other over-heated solutions.

2x2

MUNRO: THE FORMATION AND DESTRUCTION OF

Experiments on the Nitrifying Power

Tests made

Grains per gallon.

]

Nitrogen
as
nitrate.

Bristol Waters. Incubated at 80—85° F.

. Well 56 feet deep, in Coal-measures. End of pump tube
25 feet from surface. Water 10 feet from surface.
Well imperfectly closed with iron plate, water raised
by iron pump. Pure and clear

. The same water boiled after addition of NH,Cl......+.

. The same water, a few drops lead acetate added........

. The same water, accidentally overheated while incubating.

Well at least 40 feet deep, in Coal-measures. Completely
closed in and water raised by pump. Much polluted
with recent sewage

Well about 70 feet deep, in Coal- -measures. Over 50 feet
to surface of water. Open, and water raised by bucket.
Contains oxidised sewage. .

Water of old level driven under Kingswood Hill. " Sample
taken at the outflow, foot of Warmley Hill. Forms
the water supply of the immediate neighbourhood .... 3°22 |none or trace

Bristol Waterworks Co.’s water taken from their mains. . 1°4 0°03

Rain water from excavated and bricked cistern, raised by
closed pump under cover... : 0°56 trace

Rain water, caught ina sterilised beaker er during a shower.
(Added to boiled well water A.).. son eeedeeneee trace none

|
|

Downton Waters. Incubated at 55—65° F.

River water; sample taken halfway across the Avon.
Clear, with minute flocks suspended matter

Shallow well (15 feet) in alluvial gravel. Water raised
by horse gear and pumps, and delivered at the end of
tinned lead pipe 300 feet long. Pure and clear....

The same water filtered through Lipscombe’s charcoal
filter in common use

Laboratory well, very shallow, in alluvial gravel. Water ) |
raised by closed pump inside the laboratory. Pure }! 21°3 |
and clear

Shallow well in alluvial gravel, completely closed in and
water raised by pump. Contaminated with ain 29°3
three yearsago. Since purified. Clear.. °

Shallow well in alluvial gravel. Water raised ‘by pump. }

Subject to constant contamination from farmyard }| 34:2

BOWABE .c ccccccccccccece eccece eccccccccccecens

* These unnitrified waters consumed the same amount of indigotin as at first.

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS.

of Rain, River, and Well Waters.

665

with original water.

Tests made after incubation.

| c.c. Indigo |

| consumed |
by 10 c.c.
| of water.

Ammonia

| (colour with |

| Nessler).
}

|
| Nitrite (colour
with meta-
phenylene-
diamine).

Interval
after

|| making up.

Nitrite (colour
with meta-
phenylene-

diamine).

Ammonia (colour
with Nessler test).

yellow ppt.

distinct trace |

faint trace
none

amber

trace

faint trace

0°26 permillion|

deep amber
none*
*
”

*
”

yellow
pale yellow

red
yellow

faint yellow

none*

yellow
very deep orange
very faint yellow

bright orange
yellow
deep orange
none*
faint yellow
light orange
bright yellow

bright orange
yellow

deep orange

much NH; present.

NH; gone.

much NH; present.

much NH, present.
much. less NH.

much NH; present.

much NH; present.
much less NH;.

much NH; present.

”

little NH; present.
} much NH; present.

{ndigotin used at Bristol, 1 c.c.

= 0°01236 mgrm.

at Downton 1 c.c.

= 0°035 mgrm. N,

MUNRO: THE FORMATION AND DESTRUCTION OF

Depth of colour on Depth Relative rapidity of

: of colour ng
lst testing. on 2nd testing. nitrification.

. K, well water. ™ * . K, well water.

. H, river water. 2. 12. . H, river water.

. 12, well water filtered. . Be . 12, well water filtered.

. L, well water. . K, well water.| 4. L, well water.

. Il, well water unfiltered. | 5. I1. . 11, well water unfiltered.
. J, laboratory well water. | 6. J. . J, laboratory well water.

aQOor wn

All except K were in the ascending stage of nitrite reaction at the
Ist testing, and preserved the same order at the 2nd testing; K
entered the descending stage before the 2nd testing, and hence is the
most rapidly nitrifying water of the series. The Nessler test confirms
this classification; tested again after 44 days, nitrification to nitrate
was complete in K and H, the rest still contained varying quantities
of nitrite.
The Bristol waters compare with each other as follows :—

Depth
Depth of colour on of colour Order of nitrification.
lst testing. ‘ :

on 2nd testing.

. D, level water. ,
. A, well water. :
. E, Bristol Waterworks. ‘
. B, well water. ‘

2. D, level water.
. A, pure well water.
D, level water.| 4. E, Bristol waterworks.
. C, well water.
, no colour. . F, rain water.

E. . B, polluted well water.
F

. C, well water.
. F, rain water.

A
C
B

B contained neither nitrite nor ammonia when tested again on the
39th day ; D contained nitrite but no ammonia; the remaining four
still contained ammonia and much nitrite.

The sterilised solutions, it will be seen, have all refused to nitrify
added ammonia. The rainwater caught from the clouds in a sterilised
beaker is also destitute of nitrifying power.

With this one exception, all the natural waters examined possess
the nitrifying power in a greater or less degree. The polluted well
waters appeared to be most potent, then the surface waters, and last
the pure and well protected wells. Very deep well waters may
perhaps be either free altogether from the nitrifying organism, or
contain it in such small quantities as to require a very lengthened
period of incubation. It is curious that filtration through a charcoal
filter should have increased the nitrifying power of the well water I,
yet this is undoubtedly the case.

In other experiments, I found that the river water commences to

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 667

nitrify the added ammonium chloride in 7—9 days, the commence-
ment of nitrification being marked by the presence of a trace of
nitrite ; enough nitrite to give a bright orange colour with meta-
phenylenediamine was present even on the 42nd day, but nitri-
fication to nitrate was complete on the 46th day. Reinforced by the
addition of 1 per cent. of soil, this water commenced nitrifying on
the 3rd—4th day; very little nitrite was formed, which began to
diminish after the 12th day; on the 30th day nitrification to nitrate
was complete. With much soil and a thin layer of water no nitrite,
or a mere trace, was formed.

The laboratory well water, on the other hand, has never com-
menced to nitrify before the 18th—22nd day ; much nitrite is formed,
and, as has been shown above, this water.two years ago was unable to
carry the oxidation of ammonia further than the stage of nitrite. If
150 c.c. of the boiled water be seeded with 5—10 drops of a weak
nitrifying solution, about the same period of incubation is observed
(18—22 days), so that the quantity of ferment in 150 c.c. of the un-
boiled water may be taken roughly as equal to that in a few drops of
such a nitrifying solution. These experiments were made at the ordi-
nary temperature and with 5 c.c. standard NH,Cl(=5 mgrms. NH )
to 150 c.c. of water.

In the purest well waters, there is a very slight flocculent deposit
during nitrification; there is more deposit in the rapidly nitrifying
waters. Nitrification in filtered well water, or in boiled well water
seeded with a nitrifying solution free from fermentable organic
matter, would probably furnish a deposit well suited for microscopic
observation of the nitrifying organism, called by Schloesing and
Miintz, Micrococcus nitrificans. In this country, I believe its existence
has been inferred rather than made the subject of ocular demonstration.

X.—Denitrification, and the Influence of Organic Matter on Nitrification.

As already mentioned, the first three experimental solutions for
observing nitrification in river water contained—
Nov. 21, 1883.
River water + ammonium chloride.
. oo - and Rochelle salt.
i + - Rochelle salt, and soil.

The very great nitrifying power of ordinary river water without any
addition was unsuspected by me at the time, and the additions of
Rochelle salt, Rochelle salt and soil, to 8 and y respectively, were
made with the view of favouring the process of nitrification. Great,
therefore, was my surprise on examining the waters 26 days after-
wards, to find that whereas « had almost completely nitrified the

668 MUNRO: THE FORMATION AND DESTRUCTION OF

added ammonia to nitrate, 8 and y had not commenced to nitrify, but
on the contrary every trace of nitrate originally present in y was
destroyed, and most of that in 8. «@ had remained practically clear
throughout; 8 and y had developed a slight turbidity (greatest in +)
a few days after making up. The subsequent course of these waters
was as follows :—

8, which originally contained 0°204 grain per gallon as nitrate,
contained about half this quantity 26 days after making up, and still less
after another 34 days; in 30 days more there was a barely perceptible
increase; in another month the indigo consumed had doubled, and
between this testing, April 16,-1884, and August 21, 1884, complete
conversion of the added ammonia into nitrate had taken place, about
2°8 grains per gallon of N being then present in that state. (Nitrite
was present on April 16.) The solution has remained until now
without further chapge.

The nitrate originally present in y was totally destroyed during the
first month, and there was a considerable flocculent deposit. Not
until three months later was there any perceptible renitritication, but
between’ April 16, 1884, and August 21, 1884, about 2 grains per
gallon -of--ammonia was converted into nitrite, and this has since
suffered complete conversion into nitrate.

Mr. Warington has observed in his last paper (1884) that deni-
trification precedes nitrification in certain cases. Under the head of
“ Reduction: of Nitrates by Soil,” he says: “I wish simply to call
attention to the curious fact that when soil is added to diluted urine,
or to other solutions suitable for nitrification, a destruction of the
nitrates already present precedes the commencement of nitrification.

This disappearance of the nitrates is completed ina few days. * *

The reduction of the nitrates now in question is always accompanied
by turbidity of the fluid.”

No doubt this is strictly true as regards the solutions with which
Mr. Warington experimented ; these were made from substances
like asparagine, urine, rape cake, milk, &., or consisted of
artificial solutions of ammonium salts to which an alkaline tartrate,
cane-sugar, or glucose had been added. I shall, however, show
that the presence of easily fermentable organic matter is a necessary
condition of rapid denitrification; in the absence of this, soil has
no tendency to reduce nitrates already formed, and indeed if it
had it would not be easy to understand the formation and persistence
of nitrates in arable fields and in drainage waters.

The addition of Rochelle salt to a recently nitrified solution sometimes
causes a rapid and total fermentative destruction of the nitrate.

To test this point, I took the seven artificial solutions of my first
series, which had undergone complete nitrification to nitrate, viz. :—

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. | 669

Bl. Ethylamine hydrochloride nitrified to nitrate.
B2. ” ”

Cl. Potassium thiocyanate

Fl. Gelatin

El. Urine

A9. Ammonium oxalate

Al. - chloride

On December 18, 1883, all these solutions had remained for three
or four months after complete nitrification without any further
change ; they were all in bottles or flasks about half full and closely
stoppered or corked. On this date, a crystal of Rochelle salt was
dropped into each bottle, and Al (only) was heated for an hoar in
the steam-oven after this addition. When examined again on
February 7, 1884, the nitrate had totally disappeared from all except
Al; no nitrite was present in any case, and scarcely any ammonia
except in Fl, Bl, and B2. They were, however, thick with a ropy
growth of mould, so that the nitrogen of the nitrate or of the am-
monia at first formed by its reduction, had been partly consumed in
feeding these growths. Cl contained a black deposit and smelt
strongly of sulphuretted hydrogen. A1, the sterilised solution, had
remained perfectly clear, and the nitrate was intact at this date; but

a few days after the stopper was removed for testing, turbidity of the
solution set in, followed by a very active bacterial fermentation, with
evolution of bubbles of gas. After 13 days, the solution was examined
for nitrate and nitrite, and was perfectly free from them; ammonia
was however present. Under the microscope, large, very active
bacteria were easily seen with a }" objective to be swarming in the
liquid, singly and in chains of 2—8 or 10.*

Alternation of Nitrification and Denitrification.

Fl, the gelatin solution whici contained most ammonia after
denitrification, was kept for further observation. On February 7, 1884,
shortly after denitrification, it was thick with mycelium and had a
putrid odour, but by April 7, 1854, it had become clear, the odour
had disappeared, and a considerable amount of renitrification had
taken piace. By August 14, 1804, it contained no ammonia or nitrite,
but 20 mgrms. N as nitrate, or about g of the quantity nitrified the
first time. Addition of a few crystals of sodium acetate on that date
again brought about a rapid and total destruction of this nitrate, but
by November 13, 1884, a third nitrification had commenced, which has
since made considerable progress. How far this alternation of nitri-

* Probably the Bacterium denitrificans, a or B, of Gayon and Dupetit.

670 MUNRO: THE FORMATiON AND DESTRUCTION OF

fication and denitrification could be carried, I do nut know, but it is
noteworthy that the available nitrogen diminishes after each reduc-
tion, most of it being evolved as nitrogen gas. This experiment well
illustrates the relation of fermentable organic matter to denitrification.

As to the ferments themselves which bring about destruction of
nitrates, they may exist in the soil, or the water, or may be derived
from added impurity such as sewage, or even from the air. The
particular ferment encouraged will depend on the fermentable organic
matter employed, on the different species which gain access to the
liquid, and on the quality of the liquid as a nutritive medium. Since
nitrates are not only the most suitable nitrogenous food of green
plants, but are also well adapted for the nourishment of many lower
organisms, the presence of suitable organic matter in water is almost
sure to encourage the growth of some organism or other which will
sooner or later effect a destruction of nitrate. Thus a slowly growing
mould or a green alga such as that already mentioned (p. 662), may
feed on a nitrate and gradually destroy it, but this species of denitrifica-
tion is easily distinguishable from that brought about by the bacterial
ferments above mentioned. The former process may take weeks,
months, or years to completely destroy the nitrate, which disappears
only as fast as it is assimilated; the latter process is complete in a
few days, is always accompanied by turbidity of the liquid, frequently
by evolution of gas, and is marked by reduction of the nitrate to
nitrite, ammonia, or nitrogen and nitrogen oxide.

In order to exhibit more clearly the antagonism between fer-
mentable organic matter and nitrification, I will next put on record
all those bottles of the Laboratory Well Water Series (p. 7) which
received the addition of organic matter as well as of ammonium
chloride.

(In all these cases, the original N as nitrate = 0°427 grain per
gallon: the N added as NH,Cl = 2°22 grains per gallon.)

L5. Boiled after the addition of Rochelle Salt—Made up December 19,
1883. Tested on April 14, and again on August 21, not the slightest
nitrification or denitrification had taken place. The addition of
Rochelle salt to well waters causes after a time the separation of
beautifully transparent and sometimes large crystals of calcium tar-
trate. It is plain from this experiment that organic matter such as
Rochelle salt does not per se cause denitrification. Another water
boiled after the addition of Rochelle salt and soil gave the same
result.

L12. Chloroformed after addition of Rochelle Salt and Soil.—The
history of this solution is identical with that of the preceding.

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 671

Chloroform absolutely prevented either nitrification or denitrification
in the presence of soil ferments favourable to both processes.

L7. Rochelle Salt added. Not boiled or filtered.—This was contained
in a stoppered bottle, half full. The indigo consumed by 10 c.c.
at various dates was as follows :—

Dec. 19, 1883.... 1°2 ¢.c. (made up).
Jan. 24, 1884.... O09 ,,
8 eee.
Apr. 1, ,, .... 12 , Nitrite present.
o Mw scoee 36 a
Ang. 21, 1884.... No nitrite and very little ammonia. Nitri-
fication to nitrate as complete as possible.

This is an example of what I find generally takes place in good waters.
Addition of Rochelle salt produces an early and partial reduction of
the nitrate to nitrite; if ammonia is present this is presently suc-
ceeded by complete nitrification to nitrate, nitrite being first formed.

L6. Boiled after addition of Rochelle Salt; afterwards fertilised with
Soil.—The prejudicial effect of organic matter on nitrification is well
marked in this experiment, for the commencement was very slow,
and the fermentation was purely nitrous. In three months, the indigo
consumed had barely increased, but a complete conversion of ammonia
into nitrite afterwards took place, and the nitrite has persisted.
Comparing this with L3 (the conditions being precisely similar except
that one had Rochelle salt and the other had not), we see that the
soil alone would have nitrified the ammonia to nitrate long before the
commencement of even the nitrous fermentation in presence of
Rochelle salt; indeed I have met with no instance in which after the
addition of soil alone to a well water containing ammonia, the fer-
mentation has stopped at the nitrous stage.

L1l. Rochelle Salt, Soil, and a little Phenol added.—The presence
of phenol interferes with the indigo process. Nevertheless, by means
of the diphenylamine test, I have been able to ascertain that denitri-
fication was nearly or quite completed in spite of the phenol, and
that not the slightest nitrification has since taken place up to the
present day. Abundance of ammonia is still present. There is
nothing paradoxical about this. Phenol is fatal to many ferments,
and amongst others to the nitric ferment, but there undoubtedly are
bacteria and other organisms, which it is powerless in small doses
to kill.

L13. Salicylic Acid only added.—The salicylic acid was added to

672 MUNRO: THE FORMATION AND DESTRUCTION OF

this bottle as an antiseptic, in order to study its influence on nitrifica-
tion. The result was curious. Small dots or patches of mould soon
began to grow at many points on the sides and bottom of the bottle,
and in two months’ time a gradual and nearly complete denitrification
had occurred. However, after the destruction of all the salicylic
acid by the mould (shown by cessation of the ferric chloride reaction),
nitrification of the added ammonia set in and proceeded to a very
considerable extent. (Salicylic acid interferes to some extent with
the indigo process.)

Here we have a mould growing at the expense of a popular anti-
septic and destroying it, and thus paving the way for nitrification of
added ammonia.

L9. Glycerol added.—In this case, a very slow but finally complete
destruction of the nitrate occurred, with a gradually increasing floc-
culent deposit. In the course of this reduction, a little nitrite was at
one time produced. No nitrification of the added ammonia has taken

place.

L10. Crystallised Sodium Acetate added.—I was a good deal sur-
prised to find that a month after making up, this solution was totally
destitute of nitrite or nitrate. Acetate of soda, in fact, brought about
denitrification quicker and more completely than the tartrate,
sugar, or glycerol, which were tried simultaneously under identical
conditions with the same water; and up to the present date not the
slightest renitrification has taken place. Sodium acetate, in fact,
is easily fermentable in presence of a nitrate; it does not interfere
with the indigo or permanganate processes, and hence is suitable for
experiments on denitrification.

L8. Cane-sugar added.—Cane-sugar brought about a bacterial tur-
bidity, during which the nitrate was reduced to nitrite, and subse-
quently a gradually increasing flocculent deposit with total denitrifica-
tion occurred. Since then a mould has grown, but no nitrification
has taken place.

1 have now brought forward enough instances to show that what I
may call fermentable organic matter, added to a water containing a
little nitrate and also well fitted to nitrify ammonium chloride,
undergoes a fermentation which seldom fails to destroy the nitrate
already existing, and always retards nitrification of the added ammonia.
If the solutions are kept in stoppered bottles (half full), nitritication
of the ammonia may be prevented for an indefinite time, but if in open
botiles or flasks, it does eventually take place. (I have no reason to
suppose that the nitrifying ferment is destroyed durin the denitrifica-

NITRATES AND NITRITES IN ARTIFICIAL SOLUT:ONS. 673

tion, unless in exceptional cases.) It may be thought that the exclu-
sion of air from the liquid is the essential feature which determines
denitrification, and at the same time retards nitrification. This is
not the case, for rapid and complete destruction of nitrate will take
place in flasks half full and freely open to the air in the case of
polluted waters supplied with fermentable organic matter. In an
experiment on this point, when sodium acetate was added to the
water, the nitrate (1'1 grain N per gallon) was destroyed in 10 days.
A little ammonia was formed from the nitrogenous matter of the
sewage, and this nitrified during the summer.

The vigour and rapidity with which denitrification takes place after
the addition of Rochelle salt, sugar, or sodium acetate to a well water,
are closely connected with the condition of the water as regards con-
tamination. I will give an extreme instance.

From the very large quantity of nitrate and (for the district) of
chlorides contained in a certain well water, I strongly suspected the
propinquity of a cesspool, sufficiently far removed, however, to allow
of complete nitrification of all sewage finding its way into the well
by percolation through the intervening gravelly soil. At any rate the
water was condemned, and the following experiments made with
it :—

1. A crystal of Rochelle salt was added to a portion contained in a
stone jar half full. In a few days the nitrate was absolutely
destroyed, the liquid was thick with bacteria, growths of mould
rapidly followed the bacteria, not the slightest renitrification occurred
for at least eight months, and at the end of that time (August 21,
1884) the water was thick with a ropy mould, had a very peculiar
greyish-pink colour, and had developed an overpowering odour,
exactly resembling that of sewage. Ammonia was present. During
the 18 months since that date renitrification of this ammonia has
taken place, and the odour has disappeared. The jar has been corked
all the time.

2. As soon as I had noticed the very energetic denitrifying powers
of this water, I procured, on February 20, 1884, a fresh sample,
filled a 200 c.c. flask with it, added 140 mgrms. Rochelle salt, adapted
a gas delivery-tube and graduated receiver to the neck of the flask,
and placed it on a warm shelf in diffused daylight. The water
clouded on the third day, gas began to be evolved on the fourth,
increasing in amount up to the eighth, when 10 c.c. had collected.
The fermentation was now practically over, the swarms of bacteria*

* I examined these bacteria with the microscope, and made a drawing of them.
The recent publication of Gayon and Dupetit’s memoir enables me to recognise them
as Bacterium denitrificans, isolated by them in successive cultivations of a drop of
sewage in artificial media.

674 MUNRO: THE FORMATION AND DESTRUCTION OF
died down, and the water again became comparatively clear. It gave
not the slightest reaction with the diphenylamine test, but smelt
strongly of sulphuretted hydrogen, and gave a black precipitate with
lead acetate. The gas had all the characters of nitrogen.

This is the most energetic denitrification I have witnessed ina well
water not containing unoxidised sewage (no free ammonia); not only
was the nitrate totally destroyed, but the small quantity of sulphate
was reduced to H,S. It impressed me very much at the time, and
caused me to make some experiments with a view of devising a mode
of testing waters based on their power of denitrification in presence
of suitable organic matter.

Shortly afterwards the late Dr. Angus Smith’s last report appeared
in print, and in it he proposed a method of examining waters very
similar in principle. It is based on his observation that sugar, when
added to certain waters, undergoes a bacterial fermentation with
evolution of hydrogen gas. He had also observed, in 1881, that
the addition of excreta to a solution of nitre, causes an evolution of
nitrogen equivalent to the whole of the nitre employed. Gayon and
Dupetit have observed that when sewage is added to chicken broth
and other organic infusions to which nitre has been added, a strong
denitrifying fermentation is set up, resulting in the evolution of
gaseous nitrogen.

The experiment described above, and others which it is not neces-
sary to detail, led me to recognise that sewage, soils, and most
waters contain organisms which are capable of provoking denitrifica-
tion, and that the one circumstance necessary to determine this
action is the presence of fermentable organic matter—not sugar alone,
but probably the great majority of organic compounds found in plants
and animals, and including such simple substances as acetates, and
even oxalates.

Dr. Angus Smith’s test for the bacterial activity of a water was to
add cane-sugar, and observe the activity of the fermentation produced,
and the quantity of hydrogen gas evolved in a given time. He remarks
on the irregularity of some of the results as regards production of
gas, and this irregularity may, I think, be explained by the relation
of nitrates to this fermentation. The waters examined would contain
very variable quantities of nitrate, some none at all, and its
presence or absence greatly influences the result. In the first place a
water exempt from nitrate is very unfavourable to the fermentation,
because ferments of this class are greatly encouraged by a large supply
of nitric or ammoniacal food. In the second place, if nitrate is
present it may be decomposed in various ways, and may either prevent
the evolution of hydrogen or substitute that of nitrogen and oxides

of nitrogen.

SS = me @ uw. ik

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 675

It is possible, therefore, that an improvement on Dr. Angus Smith’s
proposed mode of examining waters may be made by adding a little
nitrate to waters deficient in it, together with appropriate organic
matter, and observing the extent and kind of denitrification produced.
Denitrification to a greater or less extent will always occur, whereas
evolution of hydrogen is an exceptional circumstance. I have worked
a little in this direction and will summarise my results.

The addition of Rochelle salt, sugar, acetate of soda, or even an
oxalate, to any natural water, is followed after 3 or 4 days by a
perceptible clouding, and if the water contained nitrate, this clouding
will be found by the metaphenylenediamine test to be coincident with
the production of a trace of nitrite. The air itself, if it gain access
to the water, invariably brings germs which produce this incipient
reduction. The cloudiness produced in good waters will be longer
delayed than that in polluted waters, will be much slighter, and will
disappear after a few days, leaving a very slight deposit of dead
bacteria. Unless a very small quantity of nitrate was present the
redaction to nitrite will not be complete, and the nitrite will eventu-
ally renitrify to nitrate. Unless the entering air be carefully filtered,
however, a growth of mould may be set up which will hinder this
reoxidation and even produce a further reduction.

With more doubtful waters, especially those which contain several
grains per gallon of added or original nitrate, the denitrification will
proceed farther. Turbidity will be produced in 2 or 3 days, and
coincident with this appearance will be that of a trace of nitrite. The
nitrite, however, will rapidly increase, and in the course of 3 or 4
days more nearly all the nitrate may be reduced to nitrite, so that
10 c.c. of the water will decolorise several cubic centimetres of
permanganate, and the metaphenylenediamine test will give an
immediate deep orange colour and then a precipitate. During this
fermentation, which takes place whether the bottles be full or half
full, open or corked, little or no gas will be evolved, and not a trace
of ammonia will be produced. Oxalates seem to be equally efficacious
with the other substances named in producing this purely nitrous
reduction of nitrate; at first, of course, a precipitate of calcium
oxalate is caused, but the water is perfectly clear the next day, and
gives no reaction for nitrite; on the 3rd or 4th day a fresh turbidity
is seen and is signalised by the appearance of nitrite. With these
waters the deposit is flocculent, and much more considerable than
with the best waters; but the nitrite, after its rapid formation, persists
at any rate for some time, although, if there is but little of it, it may
soon be used up at the expense of various growths. In corked bottles,
half full, a mould will generally be encouraged, and ammonia will be
afterwards found.

676 MUNRO: THE FORMATION AND DESTRUCTION OF

Other waters again quickly become clonded, and set up a fermenta-
tion which ends in total destruction of the nitrate, and is accompanied
by evolution of nitrogen gas. If the water contains 4 to 5 grains per
gallon nitrogen as nitrate, this will be completely destroyed in about 10
days from the first noticeable turbidity, and minute bubbles of gas will
escape during the entire progress of the fermentation ; an example
of th's has already been given. If the water contains very little nitrate
originally, perhaps only a few bubbles of gas will be evolved, but the
complete destraction of the nitrate instead of the production and
persistence of nitrite will mark off this fermentation from the previous
one. If, on the other hand, a considerable quantity of nitre, as well
as of fermentable organic matter, he added to the water (say as much
as 2 grams per litre of each), the fermentation and evolution of gas
will go on for, at any rate, 4 or 5 months in a suitable apparatus.
I have employed for this purpose an inverted Florence flask, full of
water, and corked, with a bent tube of fine bore passing through the
cork to allow the escape of liquid as fast as gas accumulates.

Cane-sugar, sodium acetate, and Rochelle salt, employed in
parallel experiments with the same water, all provoked this same
fermentation ; the rapidity was in the order named. When the last
two substances are used, an alkaline carbonate is the result. I cannot
say whether oxalates will support this fermentation. When the fer-
mentation is very rapid and the water contains nitrogenous organic
matter, ammonia is usually found to be present at the conclusion;
and the denitrified solution, in the case of bad waters, very easily
lends itself to rank growths of various sorts which speedily render the
whole liquid thick or ropy. Should hydrogen sulphide, however, be
produced by reduction of a sulphate during the fermentation, these
growths do not take place. The ammonia formed renitrifies after a
considerable time if the liquid be exposed to the air.

Bearing in mind Dr. Angus Smith’s observation, that excreta added
to solution of nitre produces this destructive fermentation, and
Gayon and Dupetit’s experiments, in which sewage added to nitrated
fowl broth and similar decoctions produced the same result, there is
son » presumption that a well water found to act in this energetic
manner may do so because of sewage contamination.

This is confirmed by the fact that the addition of a few drops of
fresh sewage to a good well water will cause it to take on this action.
Although the gaseous fermentation of nitre w/ll take place in bottles
only half full of water, and even in vessels fully exposed to air, yet I
have little doubt that the exclusion of air is favourable to it. Day-
light seems to exert little influence. A circumstance that may prove
fatal to such a test is the fact that, according io Dehérain and
Maquenne, all soils contain a bacterium which proyokes the butyric

PmenmawAannnaaAn..

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 677

fermentation of sugar, and in the presence of nitre a gaseous denitrifi-
cation is produced very similar to that caused by sewage. So far as
my experiments have gone, however, neither good waters nor good
waters with even the addition of a little soil, will produce this result
under ordinary circumstances. In Dehérain and Maquenne’s experi-
ments much soil appears to have been used, and they state that the
exclusion of air is essential.*

The conclusions arrived at in this paper suggest one or two remarks
bearing upon the chemistry of waters and water analysis.

In the first place, what is the nature of the organic matter contained
in potable waters ?

The soil is the abode of many ferments, some of them having
opposed functions, but all lying in wait for suitable conditions which
shall encourage one species for a little time until it has done its work
and has brought about an alteration favourable in turn to the en-
couragement of another species. From the soil, these ferments pass
into the waters, from which they are not completely removed even by
filtration, and the nitric ferment—certainly one of the most subtle of
them all—seems little affected by this process. The addition of any
ordinary organic matter instantly excites activity in one or other
of these ferments, and the effect is soon visible to the eye by the
impaired clearness of the water, and to chemical tests by the effect
prodaced on the nitrate of the water.

Now this organic matter need not be by any means that commonly
regarded as putrefiable, or even like sugar, of a nature long recognised
as easily fermentable. On the contrary, these soil and water ferments
do not spare such simple organic compounds as acetates and oxalates,
and they attack even such unlikely ones as ethylamine, cyanides,
and thiocyanates. The broadest answer then, that I can find to the
question asked above, is, that the organic matter of potable waters
can be only such organic matter as is nonfermentable, or at any rate
not rapidly or easily fermentable; and the conception of rapidly or
easily fermentable organic matter must be enlarged so as to include a
great number of substances of diverse natures. What two compounds,
for example, could exhibit a greater contrast than gelatin and
potassium thiocyanate? yet the one is as readily broken down by
soil ferments as the other. Indeed, the difficulty is to make out a list
of known compounds which are not fermentable, especially if they are
to be at the same time not fatal to organic life (antiseptics).

The following remarks on the occurrence of inorganic nitrogen in
waters apply to well waters more especially; river waters and the
water supplies of large towns represent a blend of many conditions.

* Gayon and Dupetit Lave recently shown that the hydrogen evolved in the

butyric fermentation does not reduce nitrates present in the solution.
VOL. XLIX. 22

678 MUNRO: THE FORMATION AND DESTRUCTION OF

The normal condition of an unpolluted well water I should state
thus (having regard only to the points directly raised in this paper) :
—Perfectly clear ; ammonia and nitrite absent, or present in barely
measurable quantities ; nitrate always present, but in strictly limited
amount.

Ammonia may be present as the result of a putrefactive fermentation
of nitrogenous organic matter actually in progress ; in which case the
water will be more or less cloudy. It may be persisting without sensible
alteration during the variable period of inaction of the nitrifying
ferment which follows the dying down of the putrefactive ferments—
in which case the water may be clear. It may be in process of active
nitrification, and a large quantity may disappear in a few days with a
corresponding increase in the nitrate; this is especially liable to occur
in the summer, and should a week elapse between two analyses of the
same sample of water some very striking differences in the results
would be manifest, although each analysis might be perfectly correct.
I am within the mark in stating that a well water may completely
nitrify 1 grain per gallon of ammoniacal nitrogen within a week ; but
it must be remembered that this will not take place immediately after
the addition of the ammonia. Even if the presence of the ammonia
is not associated with a putrefactive fermentation which delays nitrifi-
cation, a variable period of quiescence will take place—in fact the
rapid conversion of ammonia only takes place when nitrification has
reached its height.

Finally, ammonia may be present in the stagnant waters supporting
confervoid growths, as a bye-product of the reduction of nitrate by
various organisms. Ammonia formed by reduction is not of frequent
occurrence in well water, however, unless it is accompanied by
ammonia resulting from putrefaction.

Nitrite.—“ Nitrogen as nitrite and nitrate” has long held a place
in water analysts’ reports, but nitrite alone is seldom tested for, and
still less frequently estimated. Although nitrite is very easily formed
by both oxidising and reducing fermentations, it is very rarely present
in natural waters except in very minute traces. It will sometimes
persist in quantity, and for long periods, in artificially fermented solu-
tions and in water contained in vessels, but in contact with any large
quantity of fresh soil itis very rapidly oxidised. It may exist in a water
because the conditions do not favour complete nitrification of free
ammonia, and in this case the water may be clear; or because of a
bacterial reduction of the nitrate, caused by an influx of almost any
organic matter; in this case the water is not clear.

The reason why nitrite formed by reduction is not often found in
well waters is that in most cases the organic matter provoking the
reduction consists of sewage, and, as we have seen, sewage contains

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 679

bacteria which speedily destroy both nitrate and nitrite, with libera-
tion of nitrogen gas.

If well waters are kept out of contact with soil, nothing is easier
than to produce nitrite by nitrification of artificially added ammonia ;
but the low temperature, extreme dilution of the ammonia, and
general contact with soil and air, combine to produce a purely nitric
fermentation in most wells and rivers.

An experiment with the recently polluted well water B (p. 664),
illustrates this point. This water contained sufficient free ammonia
to give an orange precipitate with the Nessler test; about 300 c.c. of
it was therefore placed in a plugged flask, without any added ammo-
nium chloride, in order to compare the nitrification of the sewage
ammonia under the artificial conditions of these experiments with that
which takes place in situ in the well. Placed in the flask on May 24,
1886, and incubated with the rest of the Bristol waters at 80—85° F.,
the free ammonia disappeared in less than 22 days, but enough
nitrite was formed to give a bright yellow or orange reaction with
metaphenylenediamine; 10 days later the nitrite also had disappeared.
In the well itself, the water never developed sufficient nitrite to give
a recognisable colour with metaphenylenediamine.

Nitrate-—An excessive quantity of nitrate in water is very generally
regarded with suspicion; I am not aware, however, that the absence of
nitrate has been pointed out as a ground of condemnation. Clean
rain water and the water of mountain streams often contain but a
trace of nitrate; well and river waters must however contain more
than a trace unless some cause has brought about the destruction of
previously existing nitrate. This cause is the access of fermentable
organic matter to the water, and in most cases the fermentable
organic matter is derived from sewage. When, therefore, a water
contains enough mineral matter to demonstrate its percolation through
soil, and at the same time is free from nitrate, or contains only a
trace barely recognisable by diphenylamine, the occurrence of a
destructive fermentation may be inferred. These cases are not un-
common amongst well waters, and the water is generally not per-
fectly clear—the well water B (p. 664) is an example.

XI. Period of Incubation of the Nitrifying Organism.

In all published experiments, including my own, a period of
apparent inaction follows the addition of a little soil or nitrifying
liquid to a solution prepared for nitrification ; this period varies from a
few days to weeks or even months, according to the strength of the am-
moniacal solution, the quantity of seed used, and other circumstances.
At one time I thought that by excluding the tartrate or other fer-

222

680 MUNRO: THE FORMATION AND DESTRUCTION OF

mentable organic matter which encourages the rival denitrifying
ferments, I should be able to suppress the incubating period alto-
gether. This can actually be done provided enough seed be used ; but
to get sufficient or sufficiently active seed it is generally necessary to
resort to a first cultivation, during which a true incubating period,
or period of growth of the nitrifying organism, is observed.
The following examples will explain my meaning. Although the
well and river waters used were free from fermentable organic
matter, it will be remembered that 150 c.c. laboratory well water
required 21 days of incubation before it commenced to nitrify 5 mgrms.
NH;; river water required 7—9 days ;* river water + 1 per cent. of
soil required 4—5 days; and even a solution containing 150 c.c. dis-
tilled water, 5 c.c. NH,Cl, a few milligrams of K,;PQ,, and 5 per cent.
of soil, only commenced nitrifying after 3—4 days. These are periods
of true incubation: the less nitrifying ferment there is present, the
longer does it take to multiply to an extent sufficient to attack the
relatively enormous proportion of ammonia presented to it. But once
the ferment is sufficiently developed, it will produce nitrification in a
fresh solution without incubation, provided fermentable organic matter
be absent (or perhaps if all reducing organisms can be rigvrously
excluded). When the nitrified solution last-mentioned, for instance, was
poured off the 5 grams of soil and replaced by 150 c.c. fresh NH,Cl of
the same strength, nitrification commenced in less than 24 hours, and
lasted 20 days instead of 30. When this was over, the solution was
poured away, the wet soil divided into two equal portions, and each of
them covered with 50 c.c. of NH,Cl solution of half the former
strength. The bottles were also heated to 80—85°. To one of them
a little sterilised solution of Rochelle salt was added. The one without
tartrate commenced nitrifying in less than seven hours, and finished in
nine days. Here the suppression of incubation was complete. The
tartrated solution, on the other hand, although it showed a trace of
nitrite in seven hours, was completely free from nitrite and nitrate on
the third day, and nitrification of the added ammonia did not com-
mence for some days later. This represents a period of false incu-
bation; the trace of nitrite at first found proceeded from reduction
of the trace of nitrate left adhering to the soil from the previous
nitrification, and on the third day this reduction had ended in
destruction.

The commencement of nitrification in these experiments was taken
to be coincident with the first recognisable trace of nitrite (meta-
phenylenediamine test). I proved by daily testings in a separate

* Thorough aération of the river water, caused by allowing it to fall through the
air in a minute stream several times every day, did not perceptibly shorten this
period of incubation.

NITRATES AND NITRITES IN ARTIFICIAL SOLUTIONS. 681

solution that soil alone causes no redaction of nitrate to nitrite; so
that the nitrite observed, except in the tartrated solution, really
arose from oxidation of the added ammonia.

Even strong solutions of ammonium chloride will commence
nitrifying without any appreciable period of incubation if the condi-
tions indicated above are fulfilled. Thus 700 c.c. distilled water
containing 267°5 mgrms. ammonium chloride with no addition but
500 mgrms. calcium carbonate and 333 mgrms. washed soil taken from
a recently nitrified solution, in three days developed enough nitrite
for easy detection.

Appendiz.—In a memoir published since the above paper was in
manuscript, MM. Gayon and Dupetit (Recherches sur la réduction des
nitrates par les infiniment petits, Nancy, 1886) examine certain cases
of denitrification in an exhaustive manner. After recognising that
many different species of microbes will effect the reduction of nitrate
to nitrite, and that the presence of fermentable organic matter of any
description is the circumstance which determines denitrification under
ordinary conditions, they devote the remainder of the treatise to a
minute study of two species of microbes, Bacteriwm denitrificans,
a and 8, which reduce nitrates with the liberation of nitrogen gas, and
sometimes nitrous oxide. Both of these organisms were isolated for
study by successive cultivations of a drop of sewage in artificial media.
In liquids free from nitrate, they behave as aérobic bacteria, and
multiply only on the surface of the liquid, forming a zoogloa imper-
meable to air. It is even possible to cultivate them in liquids con-
taining nitrate without any destruction of the latter, provided a
continuous and thorough aération of the liquid is maintained. But
in nitrated liquids protected from the air, or only partially exposed to
it, they live at the expense of the oxygen of the nitrate, and effect the
complete oxidation to carbon dioxide of the carbon contained in the
organic matter introduced into the liquid. And it is only when this
organic matter is nitrogenous that any ammonia is formed during
the process. Nitrite is at first formed, but is speedily destroyed.
By calculation and experiment, MM. Gayon and Dupetit show that
the heat produced by oxidation of the organic carbon, minus that
absorbed by destruction of the nitrate, is a positive quantity, suffi-
cient in fact to raise the temperature of the fermenting liquid several
degrees above that of the surrounding air.

LX.—Detection and Estimation of Iodine, Bromine, and Chlorine.

By M. Dercuan, F.C.S., Lecturer on Chemistry, &c., School of Science,
Hawick.

THE various methods which have been proposed for the qualitative
determination of iodine, bromine, and chlorine, in the presence of
each other, are either so complicated or so uncertain, that any
simplication of the process of working, which would at the same
time yield more accurate results, must be of interest to the analyst.
The detection of iodine in the presence of bromine and chlorine has
been shown by Cook (Trans., 1885, 471) to be readily accomplished.
The presence, however, of small quantities of bromine and chlorine is
admitted to be much more difficult of detection. For quantitative
purposes, the separation of the halogens isa problem of even still
greater difficulty, and is one which up to the present has been but
imperfectly solved.

A strong solution of potassic dichromate at the ordinary tempera-
ture has little or no action on potassic iodide; if, however, the tem-
perature be raised to the boiling point of the liquid, the iodide is
completely decomposed in accordance with the equation 5K,Cr,0; +
6KI = 31, + 8K.CrO, + Cr,0;, and the whole of the iodine is
liberated, whereas bromides and chlorides are not in the least
affected. On adding dilute sulphuric acid, and again distilling, the
bromine is alone liberated. A simple process of separating the
halogens may be based on these facts.

The solution of potassic dichromate which I have found to yield
the best results, is prepared by dissolving 40 grams of the salt in
100 c.c. of water. A solution of this strength is without any effect
on bromides or chlorides, but the iodides are rapidly decomposed by
it. For the purpose of decomposing the bromides and liberating the
bromine, a dilute solution of sulphuric acid is added to the dichro-
mate solution. The acid solution is composed of equal parts by
volume of acid (1°84 sp. gr.) and water, and should be added to the
dichromate solution in the proportion of 8 c.c. of the dilute acid to
every 100 c.c. of water originally taken to prepare the dichromate
solution. The chlorides are partially decomposed by the acid solution
with the formation of chloro-chromic anhydride, CrO,Cl,; this, how-
ever, is not carried over with the distillate in solutions of the above
strength.

For qualitative analysis, the following method has been adopted :
a medium-sized boiling tube is fitted with a delivery-tube, bent twice

DECHAN: DETECTION AND ESTIMATION OF IODINE, ETC. 683

at right angles, one of the limbs of the tube being long enough to
reach to the bottom of an ordinary test-tube. The potassic dichromate
and water are placed in the tube in the proportions already indicated,
together with the substance to be analysed. The contents of the tube
are heated, and when ebullition commences, the end of the delivery-
tube is dipped beneath the surface of a little water, to which has been
added a drop of carbon bisulphide, contained in an ordinary test-tube :
the merest trace of iodine is thus shown. The boiling is continued
so long as the escaping vapours turn a drop of starch solution blue.
Should the boiling be prolonged, care must be taken not to allow the
liquid to become concentrated below two-thirds of its original bulk.
When the vapours no longer affect the starch solution, the tube is
withdrawn from the flame, and a quantity of the dilute acid added in
the proportions already stated. The contents of the tube are again
boiled, and the escaping vapours tested for bromine by means of a
little chloroform, or a solution of potassic iodide and starch. When
all the bromine has been evolved, the contents of the boiling tube are
rinsed into a clean beaker, diluted with water and tested in the ordi-
nary way for chlorine. For qualitative purposes care must be taken
to employ pure potassic dichromate ; this ought to be specially tested
for chlorine before being used.

For quantitative determinations the following arrangements must
be provided for: prolonged boiling, for about one hour, with some
means of keeping the volume of liquid at not less than two-thirds of
its original bulk without requiring to take the apparatus apart. The
reason why the liquid must not be concentrated beyond the point
indicated, is, that in liberating the bromine with the aid of the dilute
acid traces of chlorochromic anhydride are given off if the liquid
becomes too concentrated, and this coming into contact with the
potassic iodide solution liberates some of the iodine.

The accompanying sketch shows the arrangement of the apparatus
used for quantitative determinations :—

The flask A has a capacity of 150 c.c., the separating funnel F
serves for keeping up the volume of liquid, and also for adding the
dilute acid. The flask B attached to the lower end of the condenser
C contains the potassic iodide solution for the purpose of dissolving
the liberated iodine. It has a capacity of about 100 c.c.

The details of the process are as follows: 100 c.c. of water, 40
grams of the potassic dichromate, and about 0°4 gram of the sub-
stance to be analysed are placed in the flask’ A; the apparatus is then
put together as shown in the sketch, the contents of A heated to
boiling, and the iodine which distils over is received in the flask B.
After boiling for 10 minutes, the flask B may be removed, and a test-
tube containing a solution of starch inserted inits place. The boiling

684 DECHAN: DETECTION AND ESTIMATION OF IODINE, ETC.

is continued until a drop of starch solution is no longer turned blue
by the condensed vapours. When this occurs, the burner is removed
for a moment, and 8 c.c. of the dilute acid added by means of the
separating funnel. The contents of B, together with the other parts
of the iodine distillate, are rinsed into a beaker, and the iodine deter-
mined by means of decinormal thiosulphate solution.

The flask B is now charged with a fresh solution of potassic
iodide, attached to the condenser as before, and the distillation con-
tinued. When all the bromine is driven over—which is known by the
condensed vapours no longer turning a solution of potassic iodide and
starch blue—the quantity of iodine liberated by the bromine is esti-
mated as before, and the equivalent amount of bromine determined by
calculation.

The contents of A are rinsed into a beaker and diluted with water,
which decomposes the chlorochromic acid formed into chromic acid
and hydrochloric acid. Nitric acid is now added, and the chlorine
precipitated with argentic nitrate in the usual way.

VAPOUR-DENSITIES OF CHLORAL ETHYL-ALCOHOLATE. 685

Comparing the process herein described with those already pub-
lished, it will be readily admitted to be more easy of application, and
will, I feel sure, be found to be capable of yielding more regular and
accurate results. Those who require to determine the quantity of
bromine in kelp liquors, for which no process with any pretension
to accuracy has as yet been proposed, will find the one here
described both expeditious and trustworthy.

The following results obtained in the analysis of three mixtures
will give some indication as to the capabilities of the process.

The iodine, bromine, and chlorine were present in each case as
iodides, bromides, and chlorides, and the quantity of the halogen
present in each salt was determined gravimetrically before preparing
the mixtures :—

Chlorine. Bromine. Iodine.
No. of

mixture.

Taken. | Found. Taken. Found. Taken. Found.

0°0123 0 °0122 0 -0126 | 0°01254 | 0°01443 | 0°01441
0° 056 _ 0°0252 0 °0250 0°0288 0 02833
0°194 _ 0 -0504 0°05009 | 0:°0576 0 -05628

The chlorine in mixture II and III was not determined ; and if the
apparatus had been constructed so as to do away with the use of cork
in the fittings, the results would, I feel sure, have been nearer those
demanded by theory.

LXI.—Note on the Vapour-densities of Chloral Ethyl-alcoholate.

By Wituuam Ramsay, Ph.D., and Sypyey Youne, D.Sc.

In a memoir on “ Evaporation and Dissociation” shortly to be pub-
lished in the Transactions of the Royal Society, the results of deter-
minations of the temperatures of volatilisation, vapour-pressures (or
dissociation-pressures), and vapour-densities of chloral ethyl-alco-
holate are described, and also similar constants for chloral hydrate
and chloral methyl-alcoholate. These results appeared to warrant the
conclusion that the alcoholates are more stable than the hydrates, and
that some light might be thrown on the effects of temperature and

686 RAMSAY AND YOUNG: VAPOUR-DENSITIES OF
pressure on the dissociation of gaseous compounds by a more extended
study of the vapour-densities of these compounds.

Although it has been found that the dissociation of the vapour of
chloral ethyl-alcoholate under the possible conditions of experiment
is never very far from complete, and that it is impossible to obtain
very concordant observations, yet we think a short account of the
experimental results may be of interest.

The vapour-density determinations were made ina Hofmann’s appa-
ratus, modified, as described in a memoir on “ The Thermal Constants
of Ether” recently communicated to the Royal Society, so as to
allow of pressure, volume, and temperature being altered: at will.

The results are as follows :—

Weight of Substance = 0°01200 gram.
Pressure of Alcohol Vapour 220 mm. (reduced to 0°); Temp. 50°

(air thermometer).

Vapour-

Vapour-
density.

: Pressure. | Volume.
density.

Pressure. | Volume.

|
stile | |
|

49 °23
50°20
50°98
51°43

54°98
61°20

| condensed.

Pressure of Alcohol Vapour 350°3 mm.

; Temp. 60°.

Pressure.

Volume.

Vapour-
density.

Pressure.

Volume. |

Vapour
density.

15°05 mm.
22°0 =,
32°05 ,,

165 *55 c.c.
ot a

Re
ed ”

49-83
50°50
51°31

37°15 mm.
42°05 ,,
46 *1

oror cr
ont =
Dann
awe

Pressure of Alcohol Vapour 541°2 mm. ;

Pressure.

Volume.

Vapour-
density.

Pressure.

48°25
48°91
49 68
50°96
51°16

23°95 ,,
condensed.

CHLORAL ETHYL-ALCOHOLATE,

Pressure of Alcohol Vapour 755°65 mm.; Temp. 78°15°.

Pressure. |

Volume.

Vapour-

density.

Pressure.

Volume.

Vapour-
density.

16°15 mm,

164°9 c.c.

49°16

20°95 mm.

127°0 c.c.

49 -09

Pressure of Alcohol Vapour 747°35 mm.; Temp. 77°9°.

ure

Vapour-
density.

Pressure.

Volume.

Vapour-
density.

164°85 c.c.
100°65 ,,

Volume.

47°:
48 *5o

5 mm.

54°

49°76

Pressure of Alcohol Vapour 745°5 mm.; Temp. 77°8°.

Pressure.

Volume.

Vapour-
density.

71°65 mm.

35°05 c.c.

52°10

Pressure of Alcohol Vapour 753°2 mm. ; Temp. 78°1°.

Pressure.

Volume.

Vapour-
density.

Pressure.

Volume.

Vapour-
density.

72-2 mm.
94°35 ,,

118 °65 mm.

54°61

Weight of Substance = 0°07088 gram.

Pressure of Chlorobenzene Vapour 13415 mm.; Temp. 78°.

Pressure.

Volume.

Vapour-
density.

Pressure.

Volume.

Vapour-
density.

91°1 mm.
105°85 ,,

164 °5 c.e.
139°6 ,,

51°59
52°31

112°95 mm.
ims «

130 °0 e.c.
zis

52 64
54°00

The following determinations are reproduced from the paper
referred to :—

RAMSAY AND YOUNG: VAPOUR-DENSITIES OF

Weight of Substance = 0°02066; Temp. 78°35°.

Vapour-
density.

Pressure.

Volume.

38° 42 c.c.
36°28 ,,

52°88
52°82

122°9 mm.

34°68 c.c.

Weight of Substance = 0°07088 gram.
Pressure of Chlorobenzene Vapour 208°35 mm.; Temp. 90°.

Pressure.

Volume.

Vapour-
density.

Pressure.

Volume.

Vapour-
density.

50-06
50°99
51°27
51°73

185°55 mm.
214°5 =,
224°6 —,,

(condensed)

51°97
53 04
53°83

Pressure of Chlorobenzene Vapour 292°75 mm. ; Temp. 100°.

Pressure.

Volume.

Vapour-
density.

|
Pressure. |

Volume.

Vapour-
density.

138°2 mm.
19605 _,,
263 85,

119°2 c.e.
82°2 ,,
60°0 ,,

49 86
50°97
51°89

334°6 mm. |

355°05_,,

46° 4 c.c.
43:1 ,,

52°91
53°81

Pressure of Chlorobenzene Vapour 402°55 mm.; Temp. 110°.

Pressure.

Volume.

Vapour-
density.

Pressure.

Volume.

Vapour-
density.

158 °6 c.c. |
120°0 ,,
83°0 ,,
SO7 »

49 *89
49 *90
50°48
51°25

461-95 mm.
530°4
578°2

34°8 c.c,

29°3 ,, |

24°5 ,,
(condensed)

52 °4
54°24
59°54

CHLORAL ETHYL-ALCOHOLATE. 689

Pressure of Chlorobenzene Vapour 754°5 mm.; Temp. 131°8°.

Vapour-

Vapour-
density.

Pressure. Volume. | density.

Pressure. Volume.

— ———EE

114-05 mm. 0 c.c. 48 ‘67 301°6 mm.| 59°55cc. | 49°64
143°45 ,, . 49-01 428°5 ,, 41‘5 ,, 50°12
209°55 ,, . 49°13 595°75 29°5 ,, 50°72
295-65 _,, . 49°36 -- — —

The calculated vapour-density is 96°56. If dissociation were com-
plete, it would therefore be 48°28.

Besides the ordinary corrections for the vapour pressure of mer-
cury and the heated column of mercury, the small quantity of air in
the tube was estimated and allowed for in each calculation.

That the results are not so concordant as with stable substances is
probably to be accounted for by the fact, that when volume or tem-
perature is altered, equilibrium is not at once established. After
increasing the volume or raising the temperature, the first readings
of pressure are invariably too low, and the vapour-densities conse-
quently too high; observations were therefore made from time to
time until no further change could be detected. When the volume
was decreased or the temperature lowered, the reverse effect was
observed.

The results were plotted on curve paper, but since the total varia-
tion of density was only from 48 to about 54°5, and since the dif-
ference in the observed vapour-densities under the same conditions of
pressure and temperature amounted occasionally to more than a unit,
it was impossible to determine satisfactorily the position of the
curves representing the relation of vapour-density to pressure and
temperature. It is, however, obvious from the results that at con-
stant pressure dissociation increases, and the vapour-density is there-
fore lowered by rise of temperature, whilst at constant temperature a
similar effect is produced by lowering the pressure.

It was also found impossible to discover with certainty the altera-
tion of the density of the saturated vapour with fall of temperature,
which was readily ascertained in the case of alcohol, ether, and even of
acetic acid by drawing a curve to pass through the points of intersec-
tion of the isothermal lines with the horizontal lines representing
vapour pressures or of the isobars with the horizontal lines denoting
boiling points or temperatures of volatilisation. The chief cause of this
difficulty is the fact, noticed to a less extent with acetic acid, but not at
all with alcohol and ether, that when volume is decreased condensa-
tion begins before the pressure has reached its maximum. It may be

690 NICOL: WATER OF ORYSTALLISATION.

stated, however, that between 50° and 100° the density of the satu-
rated vapour ranges between 52° and 55°, and that there is no appear-
ance of a rise of density with fall of temperature, but that, on the
contrary, the density of the saturated vapour appears to be distinctly
lower at 50° than at 70°. Calculating the percentage number of
molecules decomposed by means of the equation

D—d
=z 100 | —_.___ },
- Cx)

where D is the calculated and d the found density, it appears that in
no case are there fewer than 75 per cent. of the total number of
molecules of chloral ethyl-alcoholate decomposed.

LXII.— Water of Crystallisation.

By W. W. J. Nicot, M.A., D.Sc., F.R.S.E., Lecturer on Chemistry,

Mason College, Birmingham.

A PAPER with the above title appears on pp. 411—432 of this volume
of the Society’s Transactions; and in it Professor Pickering endeavours
to show that experiment does not really warrant the generally
accepted view that in the heptahydrated magnesian sulphates one of
the water molecules (water of constitution) stands in‘a relation to
the salt different from the other six. In the course of the paper,
Professor Pickering adversely criticises the deductions made by me
from experiments on the molecular volume of dissolved salts (Phil.
Mag., 1883, 121; 1884, 179). These deductions were :—

That water of crystallisation occupies the same volume as solvent
water when the salt is dissolved ; but on the other hand water of con-
stitution has a molecular volume markedly different from that of
solvent water, consequently it is probable that water of crystallisation
does not exist in solution, and that water of constitution does remain
attached to the salt.

The first ground of complaint is that the data are insufficient. But
my results for water of crystallisation are not only numerous, com-
prising 23 pairs of salts, but are supported by the results of
other experimenters, amongst others by the density “moduli” of
Favre and Valson, by the similar work of Bender, by the work of
Groshans based on the experiments of Schiff, Kremers, Gerlach, and
many others; and, finally, by that of Ostwald on the volume change

NICOL: WATER OF CRYSTALLISATION. 691

on neutralisation of a large number of acids; so completely is this
established that it is stated in Meyer’s Modernen Theorien der Chemie
(1884, 452, et seg.) and in Ostwald’s Handbuch (p. 387). There is
thus no doubt that the volume of a salt in solution is a value made
up of two constants, one for the metal, the other for the salt radicle.
Water of crystallisation has no place here. With one class of salts,
and only with a section of that class, has an exception been found
to the above rule. This class is that of the sulphates: so long as
sulphates containing no water of constitution are concerned, the law
holds good, there is a constant difference between their volume and
that of other salts of the same metals. But the section of the
sulphate class known as the magnesian sulphates show a marked
difference, they have uniformly a greater volume and the differences
are consequently less. An explanation of this is to be found, I still
think, in the presence of the constitutional water molecule; at least
no other explanation has been attempted by Professor Pickering, who
complains that my results are not numerous enough, and that other
classes of sulphates should have been examined. Professor Pickering
has not considered how many soluble sulphates there are. As it
happens there are at the very most only 24 available for this form
of experiment. Of these I have examined 11, the remainder are
nearly all sulphates of the rare metals.

In my previous papers I have expressed my belief that we have no
evidence at all that the salt exists in water as a hydrate, definite or
indefinite ; and I pointed out that the heat of hydration so commonly
cited as a proof has in reality nothing to do with the question. For
though there is no doubt that the act of dissolution of a dehydrated
salt consists first in the formation of a hydrate, yet that hydrate is
solid, and the thermal results do not bear on the question of what is
the fate of this hydrate on solution. I ventured to suggest that it is
possible that the solid hydrate is decomposed on solution. This Pro-
fessor Pickering says is impossible, but it does not seem to have
appeared impossible to Berthelot or Thomsen, both of whom discuss
the question ; as a matter of fact, the question has not yet been really
touched upon, all those who have discussed the point having been
misled by the heat of hydration of the solid salt. I hope, however,
in a subsequent paper, to show conclusively from thermochemical
results that water of crystallisation does not exist in solution.

Professor Pickering’s remarks on the theory of solution proposed
by me some years ago cannot be spoken of here, with the exception
of one point which I cannot pass over in silence: surely Professor
Pickering does not imagine that he has made a point against the
theory when he writes, “ It is scarcely necessary to point out that no
amount of attraction, chemical or otherwise, can of itself produce the

692 ABRAHALL: PHENYLSULPHONIC ANHYDRIDE.

least particle of heat; it is only the actions which result from this
attraction which can do so.”

The concluding paragraph of Professor Pickering’s remarks on my
work shows that he has failed to grasp the nature of my argument
regarding water of crystallisation, based on a comparison of the mole-
cular volumes of the solutions; and he is wrong when he states that
the specific heat and specific volume in many cases of water of
crystallisation in the solid state is the same as that of solid water.
The specific heat in a few instances, and the specific volume in one or
two, is the same as that of solid water; in all the others it is markedly
different. I may say in conclusion, when Professor Pickering shows
that the difference in molecular volume of solid salts of two metals
with various acids is a constant quantity, no matter whether water
of crystallisation be present or not (if present its volume as solid
water being deducted), then I shall be quite ready to admit that it
is possible that my experiments on the molecular volumes of dissolved
salts may not be conclusive evidence against the presence of water of
crystallisation in solutions.

LXIII.—Phenylsulphonic Anhydride.

By J. L. H. Apranatt, Magdalen College, Oxford.

In the Annalen (223, 238), Hiibner describes a series of endeavours
to prepare the anhydride of phenylsulphonic acid. His last attempt,
in which he heated silver phenylsulphonate with phenylsulphonic
chloride, was partly successful, but he lost the result of his labours
by extracting the product with ether, which combined with the
anhydride to form ethylic phenylsulphonate.

In ignorance of Hiibner’s work, Professor Odling suggested the
preparation of this substance by the method mentioned above, and its
extraction with chloroform. About 30 grams of the chloride, which,
as Mr. Gossage has found, can be easily obtained in a state of purity
(m. p. 18°5°) by distilling the washed crude product under reduced
pressure (8 mm. of mercury), was mixed with a slight excess of the
silver salt prepared from sodium phenylsulphonate and silver nitrate,
the two substances having been previously dried with great care.
The mixture was heated at 160—180° for 20 hours, and the greyish-
brown solid formed was extracted with chloroform. On evaporating
the brown solution, it deposited fine prismatic crystals, which were

SENIER: ACTION OF HEXABROMACETONE ON UREA. 693

obtained almost colourless by recrystallisation. After crushing
them between porous plates the sulphur was estimated by Carius’
method :—

I. 0°2032 gram substance yielded 0°3114 gram barium sulphate,
II. 0°1890 ” - 0°2761 " ”

corresponding with 21°07 and 20°63 per cent. sulphur, the theo-
retical percentage for (C,H;SO,).0 being 21°48.

As the substance is extremely deliquescent, it was impossible to
weigh it without some absorption of water from the air. This will
probably account in part for the somewhat low percentage of
sulphur.

It melts at 54°, and when strongly heated a viscid liquid distils
over, which deposits crystals on cooling; the greater part, however,
is decomposed with formation of a black tarry matter.

It combines immediately with water, and gradually, as Hiibner has
shown, with ether. With phenol it forms a viscid liquid which I
have not yet further examined.

It is remarkable that this anhydride corresponds with sulphur
trioxide much more closely than the sulphonic chloride does with the
chlorides of sulphuric acid, phenylsulphonic chloride being inappre-

ciably acted on by water and only slowly decomposed by dilute
alkalis. In this respect it contrasts with the anhydrides of the
carboxylic acids, which are more stable in presence of water than their
corresponding chlorides.

I must gratefully acknowledge the help which I have received
from Mr. A. M. Gossage, of Magdalen College, in the preparation of
this substance.

LXIV.—On the Action of Hexabromacetone on Urea.

By A.rrep SEnIER.

SEVERAL years ago, by acting on urea with hexabromacetone, Herzig
(Ber.,12, 170) obtained two acids which he named respectively a- and
B-cyanuric acid. In many respects both these compounds behave like
ordinary cyanuric acid, but certain differences were observed which it
was thought indicated that they were distinct isomeric substances.
The attention of chemists has recently been directed to the question
of the constitution of ordinary cyanuric acid, as to whether it is to be
regarded as iso or normal, and it seemed at least possible that a
VOL. XLIX. 3A

694 SENIER ON THE ACTION OF

further study of Herzig’s acids might furnish the desired standpoint
from which to elucidate this problem. With this end in view, I
repeated Herzig’s experiments, and obtained what seemed to be two
cyanuric acids in the manner he has described. When a mixture of
urea and hexabromacetone is heated to a temperature above 150°,
bromoform and bromide of ammonium are formed, and when the
mixture is heated further to 160° or to 170—180° it then yields to
water in the one case the 8-, and in the other the a-cyanuric acid.

During the preparation of these acids, however, I was led to doubt
whether they really differed from ordinary cyanuric acid. There is
no very notable dissimilarity between the method employed by Herzig
and that by which the ordinary acid is prepared. All three acids are
formed by the action of heat on urea. The temperature, too, which
Wohler observed as that corresponding to the conversion of urea
itself into ordinary cyanuric acid, 150—170”, includes the temperatures
made use of by Herzig.

It is well known that if chlorine is used as suggested by Wiirtz
(Annalen, 64, 307), or hydrochloric acid as employed by De Vrij
(Annalen, 61, 248), the ammonia is removed more readily, being
eliminated in the one case as nitrogen and hydrogen chloride, and in
the other as ammonium chloride. In Herzig’s experiment, hexa-

bromacetone seems to act in an analogous manner to these reagents,
and to remove the ammonia as tribromacetamide according to a
reaction observed by Weidel and Gruber (Ber., 10, 1145), thus—

CBr-CO-CBr, + NH; = CBr,-CO-NH, + CHBr,.

No doubt it may be conceived that the hexabromacetone acts in
another way, forming bromoform and cyanuric acid, thus—

coc’B™ . NH, = 2CHBr, + CONH,
CBr,

but of this there is no experimental evidence, and even if it is the
fact, the resulting acid would be the ordinary cyanuric acid.

Let us now see whether a careful comparison of the properties said
to distinguish Herzig’s acid with the corresponding properties of
ordinary cyanuric acid, causes the doubt which already arises to dis-
appear, or whether as we proceed it becomes more and more defined
and rises to certainty.

The properties of the a-acid on which Herzig chiefly relies as
differentiating it from ordinary cyanuric acid are :—I. The solubility
in alcohol. II. The amount of water of crystallisation. III. The
composition of some of its salts.

HEXABROMACETONE ON UREA.

I. Solubility in Alcohol.

Herzig finds the solubility of ordinary cyanuric acid in 100 parts
of alcohol at 22° to be 0°349 part, and that of his a-acid to be 0°556
part. In other words, the «-acid is about one and a-half times as
soluble as ordinary cyanuric acid. It is remarkable that, notwith-
standing this small difference, only a single determination should have
been made. My experiments with ordinary cyanuric acid leave little
room for doubt that its solubility in alcohol is influenced considerably
by small variations in the experimental conditions.

Six determinations made at temperatures very close to that of
Herzig’s experiment, give me the following numbers which are seen to
exhibit considerable differences :—

L. Il. III. IV. V. VI. Mean.
0088 0:095 0°090 0090 0100 0-110 0°095

These results were obtained by allowing hot solutions to cool and
then remain with occasional agitation for one or two days. They
indicate, on the whole, that the solubility of cyanuric acid in alcohol
is less than one-third of that found by Herzig; but the numbers are
not as concordant as is usual in such estimations. This is explained
by the tendency to form supersaturated solutions in alcohol, which, as
will be seen from the next experiments, is clearly characteristic of
cyanuric acid. Hot saturated solutions were prepared, and the
amount of acid held in solution was determined from time to time.
At 22—23° 100 grams of alcohol retained in solution—

Experiment 1. Experiment I
gram. gram.

After two hours.... 0°192 After one day.... 0°122
After two days After two days .. 0082
After three days.. 0°090
After four days.. 0°075

It is seen that on standing the solution contains less and less cyanuric
acid. If we take the mean of these numbers, the solubility of cyanuric
acid in 100 parts of alcohol is 0106 part. This is, however, a
somewhat higher figure than that resulting from the previous deter-
minations (0°095). As the mean of the whole 12 experiments the
solubility of cyanuric acid in 100 parts of alcohol at 21—24° is
0100 part, whilst from one experiment Herzig concludes that its
solubility is 0°349 part. Herzig thus finds ‘the solubility to be more
than three times as great as my experiments indicate. If the relation
of my mean result for ordinary cyanuric acid to Herzig’s number, on

34a 2

696 SENIER ON THE ACTION OF

the one hand, be compared with that of the numbers found by Herzig
for the ordinary, and the a-acids on the other, thus—

0°100 : 0°349 and 0°349 : 556,

then it is evident that the ratio of difference between two observations
of the solubility of the ordinary acid is greater than that between the
ordinary and the a-acid as given by Herzig. Under these circum-
stances, a small difference in solubility of two substances can be
accepted only very reluctantly as evidence of difference of con-
stitution.

The solubility of cyanuric acid which, in accordance with Herzig’s
experiment, has found its way into chemical literature (for example,
see Beilstein, Handbuch Org. Chem., 2nd Ed., 1006), is therefore too
high. If from the twelve experiments I have made, two which are
clearly cases of supersaturation are removed, the mean of the remain-
ing ten, 0°09, may be taken as very accurately representing the
solubility of the ordinary cyanuric acid in 100 parts of alcohol
at 21—24°.

In the next place the relative solubility of the ordinary acid and
of Herzig’s a-acid must be examined.

The difference of solubility in the case of these two acids as
observed by Herzig, and as the result of two experiments of my own,
is as follows:—

Senier.
ae
Herzig. I. II.
Ordinary cyanuric acid .... 1°00 1:00 1:00
Herzig’s a-acid 2°15 1:22
The solubility of the ordinary acid is taken as unity for purposes of
comparison. In agreement with Herzig, I find the a-acid more soluble
than the ordinary acid. But although in my first experiment it is
more than twice as soluble, in my second experiment its solubility is
not much greater than that of the ordinary acid.

It was observed by Herzig that his a-acid could not be converted
into a salt and separated therefrom without being changed into the
ordinary acid. Regeneration from a barium salt gave him an acid
having the same solubility in alcohol as the ordinary acid. I have
repeated the observation and can confirm this statement, thus :—

Senier.
Ordinary cyanuric acid 1:00
a-Acid regenerated from a

barium salt 1:10

The most probable explanation of these observations is that minute
quantities of foreign substances are present which are sufficient to

HEXABROMACETONE ON UREA. 697

increase the solubility of the acid. The purer the acid the more it
approaches to the solubility of ordinary cyanuric acid. Conversion of
the acid into a salt and regeneration therefrom does not appear to be
a process cf intermolecular change, but simply a method of purifica-
tion. If this is true, then the solubility of the a-acid should be
reduced when it is purified by recrystallation, for instance, from
alcohol. Here are the results of some experiments to determine this
compared with the numbers already obtained :—

a-Acid.

Ordinary
cyanuric acid. | Crystallised Regenerated Recrystallised from
| from water. | from barium salt. alcohol.

; IL.
1°10 | 1°15 1°09

1-00 | 2°15
|

Recrystallisation from alcohol itself, then, it is evident, is enough
to so purify the a-acid that its solubility in alcohol is reduced very
closely to that of ordinary cyanuric acid.

Numerous attempts have been made to discover the impurity the
presence of which is the cause of the difference in solubility observed
in the case of the acid prepared by the use of hexabromacetone.
These inquiries have not, however, led to a perfectly satisfactory con-
clusion. It is deserving of mention, nevertheless, that when cyanuric
acid itself is recrystallised from a solution containing small quantities
of biuret, its solubility in alcohol is much increased.

Herzig’s observation that fuming nitric acid also converts the a- into
the ordinary acid is quite in accordance with the explanation to
which I have been led of the behaviour of the a-acid when regenerated
from a salt or recrystallised from alcohol, namely, that both these
processes are cases of simple purification. This very method was
employed long ago by Liebig and Wohler (Berz. Jahresbericht, 11, 83)
as a means of purifying ordinary cyanuric acid. These observers
dissolved the crude cyanuric acid in hot oil of vitriol, and to the
solution added nitric acid, afterwards precipitating the pure cyanuric
acid by water.

II. Water of Crystallisation.

The proportion of water held by crystals of ordinary cyanuric acid
was determined by Liebig and Wohler (Pogg. Ann., 20, 369).
When the acid is freshly recrystallised and pressed between paper, it
contains 21°8 per cent. of water, corresponding to two molecules.
Herzig found in the case of his 2-acid only one molecule. His results

698 SENIER ON THE ACTION OF

are 11°73 and 12°13, corresponding very well with one molecule, which
requires 12°24 per cent.

Repeating this experiment with a specimen of the a-acid which I
prepared, I found, after freshly recrystallising from water and drying
between paper, that it contained 17°6 per cent. water, which was
given off at 120°. This specimen was then recrystallised from alcohol,
and again from water, when its power of combining with water had
risen very nearly to that of the ordinary acid. It contained 19°9 per
cent. of water, whereas a determination in the case of the ordinary
acid gave me 20°6 per cent. Thus by purification it is seen that the
percentage of water of crystallisation which the acid is capable of
fixing increases until a number is reached not far removed from that
given by the ordinary acid.

III. Composition of some of its Salts.
1. Silver Salis——Herzig describes two silver salts of his a-acid. A
complete analysis of the one led him to give it the formula

(CaN) < tO NG LO,

which requires 57°14 per cent. of silver. This is probably, as Herzig
himself remarks, the same salt that Wohler (Annalen, 62, 244) pre-
pared, but to which the somewhat different formula

(CN) <(ON ES NHS
is ascribed, which is based, however, merely on a silver determination.
This formala requires 57°29 per cent. of silver.

A trisilver salt of the ordinary acid of the composition

(C:N;)(OAg)s,H20,
which requires 69°23 per cent. of silver, was prepared by Liebig
(Anunalen, 62, 123), and afterwards by Wohler (Annalen, 62, 245).
Now following the method of Liebig, Herzig obtained in the case of
the a-acid salts containing from 57 to 59 per cent. of silver, and this is
noted as characteristic of the a-acid.

I have repeated this experiment and have obtained salts with various
proportions of silver both from the a- and the ordinary acid. By the
careful employment of an excess of ammonia, according to Liebig’s
method, I obtained, however, from both acids the trisilver salt, as the
following analytical numbers indicate :—

Silver salt of Silver salt of
Theory. eyanuric acid. a-acid.

Silver .... 69°23 69°5 68°7

When the ammonia is not in excess throughout the operation, salts

HEXABROMACETONE ON UREA. 699

are produced containing a less proportion of silver. Such salts were
observed by Wohler in the case of cyanuric acid and Herzig’s salts
the a-acid probably belong to this class. One which I obtained gave
me 58°1 per cenf. of silver.

2. Barium Salts—Wohler (Annalen, 62, 251) discovered mono-
barium cyanurate, which he prepared by dropping a solution of barium
hydroxide into a boiling solution of cyanuric acid until, the still liquid
remaining acid, the first permanent cloudiness appears. The solution
is then kept at 60° for several hours, when the salt is precipitated in
beautiful crystals. With ordinary cyanuric acid, Herzig working in
this manner obtained Wohler’s salt, C;N;(Oba)(OH), + H,0.

Herzig’s analysis gave—

Experiment.

Barium | 29°72

Now the same experiment in the case of the a-acid led to a different
compound, and this is regarded by Herzig as a characteristic property
of this acid. The salt obtained contained 41°29 per cent. of barium,
corresponding to the formula C;N;(Oba),(OH) + 4H,0.

But repeating this experiment I arrived at results which are
in accord with the general outcome of my previous experiments. I
have found when operating with the a- as well as with the ordinary
acid, that the same monobarium salt is formed which Wohler first
observed. Thus I obtain the following nambers :—

a-Acid recrystallised
Theory. Cyanuric acid. from alcohol.

Barium 31°93 31°73 29°77

The percentage of silver which my a-barium salt contained is the
same as that found by Herzig in the salt of the ordinary acid.

By the use of larger proportions of barium hydroxide salts richer
in barium are easily produced. From the a-acid I have prepared
salts or mixtures of salts containing 41°4 per cent. and 50°23 per cent.
of barium, and from the ordinary acid one containing 45°4 per cent. ;
both these salts and Herzig’s salt, which gave 41°29 per cent. of
barium, are very probably mixtures.

The doubt which was felt during the preparation of this 2-acid as
to whether it is to be looked on as a distinct isomeride of ordinary
cyanuric acid or not, has gradually become stronger, and the conclusion
to which I am led by the whole investigation is that the acid produced
by the action of hexabromacetone and heat on urea and known as
Herzig’s a-cyanuric acid, is identical with ordinary cyanuric acid.

Note on B-Cyanuric Acid.—This acid is produced in the same re-
action as the a-acid, but at a somewhat lower temperature. Whilst

700 LAURIE ON THE MEASUREMENT OF

the a-acid requires a temperature rising to 180°, the 6-compound is
formed between 150—160°. It is more soluble in alcohol than the
a-acid, but Herzig gives no determination of its solubility. It crystal-
lises from water in anhydrous crystals. Further, it is distinguished
from both the a- and the ordinary acid by not giving cyanic acid when
heated, and by not forming cyanuric chloride when treated with phos-
phorus pentachloride. Finally, when dissolved in soda solution and
the solution warmed, it does not give the sparingly soluble tri-
sodium salt, which, as pointed out by Hofmann (Ber., 3, 769), is so
characteristic of ordinary cyanuric acid.

My experiments give me somewhat different results. I have
undoubtedly obtained cyanuric chloride by the action of phosphorus
pentachloride. This I have isolated from the crude product of the
reaction, and found it to melt at 144°. Moreover, it is easy with my
specimen of the A-acid to obtain the difficultly soluble trisodium salt.
That it is really the B-acid described by Herzig is proved by the fact
that it did not give cyanic acid when heated in a small tube, a
test which never fails, even with small quantities, in the case of pure
cyanuric acid. Whether, therefore, this B-acid is a distinct isomeric
compound cannot as yet be decided with certainty, although the
previous work on the a-acid renders this improbable, and a final
judgment must await further experiments.

University Laboratory,
Berlin, June, 1886.

LXV.—On the Measurement of the Electromotive Forces produced by
the combination of Cadmium and Iodine in Presence of Water.

By A. P. Lavrig, B.A., B.Sc.

In the March number of the Philosophical Magazine for this year,
I published a paper on the “ Electromotive Forces developed by the
Combination of Zinc and Iodine.” Instead of completing those
experiments, I have taken up the reaction between cadmium and iodine
on account of the greater ease with which cadmium iodide is manipu-
lated. The following is an account of my preliminary experiments
on this reaction, and the results which I think may be deduced from
them.

As is well known, cadmium and iodine combine in the pre-
sence of water to form cadmium iodide, which dissolves. Just at
the moment of combination there is a difference of electric potential

ELECTROMOTIVE FORCES.

between the cadmium and the iodine atom, which we can measure by
an arrangement which I shall presently describe. This difference of
potential is modified by the surrounding conditions. In the first
place, we may consider that it is made up of two parts, due respec-
tively to the combination of the cadmium and iodine, and to the
solution of the resulting molecule in the water.

Let us first consider that part which is due to the solution of the
resulting molecule. If we imagine the reaction to be going on, we
evidently have cadmium iodide accumulating in the water present,
until at last we may suppose that the water becomes saturated with
the salt. When this state is reached, it}is evident that there is no
further difference of potential between the salt and the water, that
this factor of the whole difference of potential has disappeared. No
doubt then this portion of the difference of potential has been dimi-
nishing from the beginning of the reaction, but it may diminish in
two ways. If on the one hand there are any definite compounds
formed between the salt and the water, we should expect on passing
through the strength of solution corresponding with such a com-
pound to have a sudden change in the difference of potential. If on
the other hand no such compounds are formed, but the difference
of potential between the salt and the water is a continuous function
of the salt already in solution, we should expect the potential to fall
off along a continuous curve, flattening as we approached a saturated
solution.

Let us now consider the other part of the reaction, namely, the
combination of the cadmium and iodine. The cadmium remains a
solid metal throughout the reaction, and we may therefore suppose it
to be uninfluenced by changes in the solution. The iodine is very
slightly soluble in pure water, but is increasingly soluble as the
amount of cadmium iodide in solution is increased. There is prob-
ably then a difference of potential between the iodine and the
cadmium iodide, this difference of potential diminishing as the
amount of iodine present in a given solution is increased, and finally
disappearing when the solution is saturated with iodine. This
difference of potential must evidently diminish the difference of
potential due to the combination of the cadmium and iodine, as the
iodine has to be torn from its solution in cadmium iodide. We may,
however, consider that when the solution is saturated with iodine, the
iodine is then free to combine with the cadmium.

We should expect then that if we increase the amount of iodine
present in a given solution, the difference of potential will increase
between the cadmium and iodine, and in this case also it will increase
in jumps if we pass through any compounds of cadmium iodide and
iodine; but if there are no such compounds, the difference of potential

702 LAURIE ON THE MEASUREMENT OF

between the cadmium and iodine will increase along a continuous
curve, the increase, however, becoming less rapid as we approach the
saturation of the solution with iodine.

We can, however, diminish this difference of potential in another
way. If we suppose a given solution to be saturated with free iodine,
and we then add more solid cadmium iodide so as to increase the
amount of cadmium iodide in solution, we shall immediately diminish
this difference of potential, because we are now no longer dealing with
a liquid saturated with iodine; the liquid is capable of dissolving a
great deal more iodine, and consequently a difference of potential has
been set up between the iodine and the cadmium iodide.

I have explained in my Zinc-Iodine paper how, by integrating such
a curve as one of those above described, we can obtain a measure of
the whole electrical energy produced by the reaction, so that I need
not enter into it here. With a view to measuring these differences of
potential I have devised the following voltaic cell.

A cadmium plate and a platinum wire are immersed in a solution of
cadmium iodide containing a little free iodine. On connecting the
cadmium and platinum plates a current passes through the cell, the
cadmium combining with the free iodine to form cadmium iodide,
which dissolves in the water present. It is hardly necessary to
explain that it is the free iodine round the platinwm wire which is
used for this combination. On connecting this cell with a Thomson
quadrant electrometer, I could measure its electromotive force against
a Latimer Clark cell, the electrometer reading to 0°005 of a volt. In
every case two cells in series were used, and the Latimer Clark has
been kindly tested for me by Professor Ayrton against his standard.
Its E.M.F. at 10° C. is 1:433 volts (the temperature at which the
experiments were made). Commercial cadmium and polished electro-
cadmium were used. They gave the same deflection on the electro-
meter.

The method employed was to make up various solutions containing
varying amounts of cadmium iodide and free iodine, and then place
them in the cell and measure the E.M.F. produced. In this way we
measure the E.M.F. produced by the combination of an infinitely
small quantity of cadmium and iodine under definite conditions, and
so determine a point on one of our curves.

A very brief description of each group of experiments will be
sufficient after the above explanation.

Firstly, I tested the effect of diminishing the amount of free iodine
present, the cadmium iodide in solution remaining the same. The
results are given in curve and Table A. The experiments were not
made so carefully as those that follow, but are sufficient to show that
the E.M.F. is a continuous function of the free iodine present, and

ELECTROMOTIVE FORCES. 703

that the curve flattens off as we approach saturation with iodine as we
should expect. There is no indication of a compound.

TaBie A.

Grams of free iodine
in 1 c.c. of a 1 per cent.

E.M.F. of cadmium cadmium iodide
cells. solution.

1-087 0:0032
1-081 0:0016
1-067 0-0008
1-063 (?) 0:0004
1-050 0:0002
1-045 0:0001
1-039 0-00005

oj
$
|
B
z
3
=
o
ose
)
fs
a
om

Milligrams of free iodine in 1 c.c. of solution.

Secondly, I tested the effect on the E.M.F. of the cell of increasing
the amount of the cadmium iodide present. Each solution was
saturated with iodine with the view of having the iodine free to
combine with the cadmium in each case, so that I could measure the
changes in the difference of potential between the cadmium iodide and
the water alone. The results are given in curve and Table B. There
isno indication of a compound between the cadmium iodide and the

704 LAURIE ON THE MEASUREMENT OF

water, but the curve is continuous, flattening as we approach
saturation.

TaBeE B.

E.M.F. of Grams of cadmium Grams of free iodine
cadmium iodide in 100 in 1 c.c. of the
cells. grams of water. solution.
1119 0-1 0°0008
1:087 1:0 0°0032
1:081 50 0°0071
1:072 11°4 0°0095
1°057 30°0 0°0113
1:046 50°0 00151
1:036 70°0 0°0202
1:034 81 sat. at 10° C. 0°0300

Thirdly, I tested the effect of increasing the amount of cadmium
iodide present, keeping the amount of iodine present constant. The
results are given in curve and Table C. Here we have two effects,
one superposed on the other. Firstly, the gradual diminution of the
difference of potential between the cadmium iodide and the water,
which is shown by curve B. Secondly, the diminution in the dif-
ference of potential between the cadmium and the iodine, due to the
effect on the iodine of the increasing amount of cadmium iodide in
solution. Evidently any number of curves, starting from different
points along curve B and lying below it, could be drawn for different
constant amounts of free iodine present. These experimental results
then bear out the description of the probable variations in E.M.F.
given at the beginning of the paper.

TaBLeE C.
E.M.F. of cadmium Grams of cadmium iodide
cells. in 100 grams of water.
1:084 01
1:060 0°49
1:045 ll
1°025 51
1:018 0:0
1-008 20°0
1001 30°0
0994 40:0
0°985 50°0
0-980 60°0
0°975 70°0
0:974 80:0

0°974 81:0

ELECTROMOTIVE FORCES.

a
~
—

—

=
—
=

eo
1S)
od

°
ey
aa

Grams of salt in 100 grams of water.

We have next to consider the effect of change of temperature on the
electromotive force of the cell. The importance of such measure-
ments has been sufficiently explained in my Zinc-lodine paper, where I

706 LAURIE ON THE MEASUREMENT OF

have given an account of Professor Helmholtz’s work on this subject.
It is sufficient to state here that we can obtain a measure of the differ-
ence between the electrical energy produced during a reaction, and the
whole heat set free by the reaction, if we measure the variations in
E.M.F. caused by variations in temperature.

The calculation is made by means of the following formula :—

de
TS = J.dQ,

where de is the variation of E.M.F. caused by altering the absorbed
temperature T by the amount dt; and dQ is the difference between
the electrical energy and the heat produced by the reaction. I have
then, applying this formula, made some experiments to settle the
following question. When a salt is dissolved in water, the heat set
free or absorbed is made up of two parts, the heat of solution of the
salt in water, and the latent heat of liquefaction of the salt. It is
usually assumed that when a salt is formed in presence of water, as in
the combination of cadmium and iodine, that this latent heat enters
into the thermal result, that the salt is formed in the solid state and
then dissolved : I have obtained some evidence that the salt is formed
in the liquid state. Iodide of zinc sets free heat on dissolving in water.
Iodide of cadmium absorbs heat on dissolving in water. Nevertheless,
both salts give similar curves as representing their E.M.F, of solution
in water. This naturally suggests the question whether the latent heat
of liquefaction enters into the energy set free when the salt is formed
in presence of water, or to put it in another way, whether the salt is
formed in the solid or formed in the liquid state, when the metal and
iodine combine.

If, as I suppose, the salt is formed in the liquid state, and the
latent heat of liquefaction does not enter into the energy set free
during the reaction, we should expect that in a cell in which solid
salt was separated by the passage of the current, there would be
a considerable rise of E.M.F. due to the heat set free by the passage
of the salt, when formed, from the liquid to the solid state. If there
was no rise in E.M.F., and the curves B and C show that there is
none, then we should have to suppose that this latent heat appeared as
heat in the cell.

This then is a point for experiment by the application of the above
formula. I have found the measurement of temperature effect very
troublesome, as the iodine rapidly attacks the cadmium at high tem-
peratures, and have had therefore to use a cadmium-silver iodide cell.
This consists of a cadmium plate and silver wire coated with silver
iodide, immersed in a cadmium iodide solution. I have used five or
six cells in series, placed inside an ordinary hot water oven. I first

ELECTROMOTIVE FORCES. 707

experimented on cells containing dilute cadmium iodide. This cell
showed no variation on the electrometer on being raised through
40° C., showing that its E.M.F. does not vary 0°00002 volt per degree
Cent. I next experimented on cells containing saturated solution of
cadmium iodide. I detected a slight diminution in E.M.F. on raising
them through 40° C., but not enough for measurement. These two
sets of experiments may be taken as showing that practically the
whole of the energy set free by the reaction between cadmium and
iodine appears as electrical energy. Thirdly, I experimented on cells
containing a large excess of solid salt, so that the solution should be
saturated throughout the whole range of temperature, and conse-
quently the passage of a current would separate solid cadmium iodide.
I immediately obtained a very considerable diminution of E.M.F. on
warming the cell. I found it impossible to measure it accurately,
as for various reasons both the E.M.F. of the cells and the tempera-
ture effect kept varying.

A very laborious series of experiments ultimately enabled me to
detect the causes of these fluctuations, but not to remedy them with
the apparatus at my disposal. I therefore only give a rough approx-
imation to this temperature effect at present. It is about 0°00037 volt
per degree C. This means that in a cell in which the passage of the
current causes the separation of solid salt there is set free about
4000 heat-units for every gram-molecule of cadmium iodide. The
value of the electrical energy set free in heat-units by the formation
of the solution CdI,,400H,0, is about 2000 heat-units. Consequently,
if the theory described above is correct, we should expect an absorp-
tion of heat of about 2000 heat-units on dissolving cadmium iodide
in water (CdI,,400H,0). This result is mach too high, as Thomsen
gives the value 1000 heat-units as the amount of heat absorbed. As
already explained, the value given for the temperature effect is a very
rough approximation, so that I think this result, as far as it goes,
distinctly supports the above explained theory that when a salt is
formed in presence of water, the latent heat of liquefaction does not
enter into the energy set free.

Before concluding this paper I wish to draw attention to a point of
some interest. On looking at curve B or C, it is obvious that a
trace of salt, as little as 01 per cent., has a marked effect on the
E.M.F. of the cell. This can hardly be explained on the usual theory
of molecular compounds, unless we are prepared to write as one
molecule CdI,,20000H.O. Some other explanation must be found,
but whatever this may be, the result is noteworthy as showing the
enormous distance, compared to its molecular diameter, to which the
molecule is able to influence the difference of potential between the
water and the next molecule of salt which is going to dissolve.

708 ANSCHUTZ AND EVANS: CHLORIDES OF ANTIMONY.

Whether or no there is any limit to this can only be decided by
experiments made with water prepared with the utmost care.
Ordinary distilled water cannot be considered as sufficiently pure,
but on trying it I obtained an E.M.F. higher than any given on
Table B.

There is another subject on which the curves probably have an
important bearing, namely, partial double decomposition between two
salts in solution, but I have not yet sufficient experimental evidence
to justify me in explaining their ir. ‘tance in this direction. In
conclusion, I must explain that all experimental details have been
omitted from this paper, but will probably be published elsewhere.

LXV1I.—Contributions to our Knowl. ,e of the Chlorides of Antimony.

By Ricuarp Anscniitz and P. Norman Evans.

In the course of a research on the action of anhydrous oxalic acid
on inorganic chlorides, we have made an observation with reference to
antimony pentachloride that appea to us remarkable.

We prepared pure antimony pentachloride by the action of chlorine
on antimony trichloride. To assure ourselves of the purity of the
trichloride it was distilled under diminished pressure, and was found
to boil constantly at 113°5° under 23 mm., and at 103° under 14 mm.
pressure.

The pure antimony trichloride was supersaturated with chlorine
and the liquid product then distilled under diminished pressure to
purify it.

Trial I—19°15 grams of SbCl, after saturation with chlorine gave
26°45 grams crude product, this after distillation gave 24°65 grams
of SbCl, instead of 25 grams reckoned.

Trial II.—126°5 grams of SbCl, after saturation with chlorine gave
174°5 grams crude product, this after distillation gave 163-1 grams of
SbCl, instead of 166°1 grams reckoned.

These figures show that antimony pentachloride boils without
sensible decomposition under diminisho. pressure, and this was further
confirmed by two determinations of the chlorine.

I. 0°2419 gram SbCl; gave 0°5796 gram AgCl.
II. 0°1720 " », 0°4120 ”
Found.

Calculated for SbCI,. A
Cli... 5924 oo" 59°26

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID. 709

It is interesting that the boiling point of antimony pentachloride
lies considerably lower than the boiling point of the trichloride under
the same pressure: thus, under 22 mw. it boils constantly at 79°, and
under 14 mm. at 68°.

As in the literature of the subject we can find no details of the
dissociation of antimony pentachloride, we have resolved to occupy
ourselves with a study of this question. We reserve a discussion as
to whether indeed antimony pentachloride boils under a low pressure
without decomposition, until yp shall have made a vapour-density
determination under much diminished pressure. Should there be no
true boiling under our conditions of pressure but rather dissocia-
tion, it is remarkable that antimony trichloride should exist as a gas
in the gaseous mixture, at a teu:perature which is considerably under
its own boiling point at the ame pressure.

Bonn University, July, 1886...

LXVII.—Pipitzahoic Acid.’ (First Communication.)
By Ricwarp Anscniitz and Joun Water LEATHER.

In the year 1855 Weld (Annalen, 95, 188) published a paper entitled
“ Ueber die Pipitzahoinsaeure, einen eigenthiimlichen Pflanzenstoff,”
describing an acid substance which Professor Rio de la Loza had
extracted from the roots “raiz del pipitzahuac” and had called
‘“pipitzahoinsaeure.” Liebig received a small quantity of this
pipitzahoinsaeure (pipitzahoic acid) from Dr. Schaffner, a German
physician, of San Luis Potosi (Mexico), and gave it to Weld for in-
vestigation. In 1856 Schaffner discovered the plant itself and gave
it the name Trizis pipitzahuac. Specimens of the plant and large
quantities of the dried roots, together with a large quantity of other
Mexican plants formerly the property of the late Dr. Schaffner, came
into the hands of his friend Mr Vigener, apothecary in Biebrich on
the Rhine. 3

Vigener again drew the atterition of the scientific world to the
raiz del pipitzahuac, which in Mexico is highly valued medicinally, the
inhabitants using it as an energetic purgative, its application being
unaccompanied with disagreeable results. On March 3, 1884, Vigener
(Sitzwngsber. d. niederrhein. Gesellschaft in Bonn, 1884, 86) laid

samples of pipitzahoic acid before a meeting of the Niederrheinische

Vul. XLIX. 3B

710 ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

Gesellschaft fiir Natur- und Heilkunde in Bonn, declaring it to be an
anthraquinone-derivative, and suggested that one of us (Anschiitz)
should submit it to a chemical investigation, a proposition which was
accepted with pleasure. The reactions of pipitzahoic acid which
were then observed by one of us (Anschiitz) (ibid., 1884, 145), led
him to form an opinion of the constitution of this acid which has
not been altered by any of our later experiments.

With regard to the anatomical structure of the plant from which
pipitzahoic acid is derived, there is probably scarcely a substance
which is so clearly seen on superficial examination as pipitzahoic acid,
as it is contained only in the pipitzahuac roots and is distinguished by
its golden-yellow colour. At our request Professor Schimper had the
kindness to examine the roots, and we take this opportunity of
rendering him our best thanks for the following communication :—

“The yellow crystalline substance of the root of Trixis pipitzahuac
is contained in large, cylindrical, intercellular passages which are
limited to the sieve tubes of the fibro-vascnlar bundles, three or four
occurring in each of the bundles. In the remaining parts not a trace
of this substance can be recognised even with the highest power of
the microscope.

** From the mode of its appearance it may be assumed as extremely
likely that pipitzahoic acid represents a product of metastasis which
is of no farther use in the plant, similarly to the ethereal oils in the
resin-canals of the Umbellifere. Similar receptacles occur among
the representatives of the Composites to which the Trizis pipitzahuac
belongs, and it is probable that the contents of such receptacles have
in all cases the same physiological value.

“A similarity in the manner of its appearance with that of the
colouring matter of the madder-root (Rubia tinctoriwm) does not
exist. Alizarin (more properly its chromogen) is not confined to
special receptacles, but is contained in all parenchymatic cells,
especially those of the bark.

“The root of Trixis pipitzahuac is somewhat rich in glucose, its
parenchyma contains colourless, sharp-edged bodies which are slightly
optically active, but do not appear to me to be crystalline. The small
intercellular spaces around the fibres, which are dispersed in small
groups in the bark and the pith, are filled with a brown solid sub-
stance.

“‘ Substances other than those above-mentioned could not be dis-
tinguished by the aid of the microscope.”

The appearance of pipitzahoic acid suggested that this substance
might be included in the group of the quinones. Regarding in the
light of our present knowledge the facts discovered by Weld years
ago, more especially the property of forming alkaline salts which

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID. 711

are decomposable by carbonic anhydride and dissolve in water with a
beautiful violet-red colour, it appeared probable that pipitzahoic acid
is a hydroxyquinone, and the comparison of pipitzahoic acid with
alizarin was also thereby suggested: but a little consideration showed
that pipitzahoic acid cannot belong to the group of the hydroxy-
anthraquinones, especially as the form in which it appears in Trizxis
pipitzahuac is quite different from that of alizarin in Rubia

tinctorium.

Taking into account both Weld’s observations and experiments of
his own which had served to determine the quinone-like character of
pipitzahoic acid, and the formula also C,H»O; (with old atomic
weights CyH»»O,), one of us (Anschiitz, Ber., 18, 709) published some
time ago the following considerations as to its constitution.

“Tf from the formula of pipitzahoic acid the formula of monhydroxy-
quinone be subtracted, there is a difference of C,Hi, = C,s;H»O; —
C;H,O;. Now supposing that in hydroxyquinone one of the three
benzene hydrogen-atoms is displaced by a single hydrocarbon-group,
this side-chain must be C,H,;, but the group C,H,, contains so many
hydrogen-atoms that only two of the nine carbon-atoms can be
doubly linked, and therefore a second benzene-ring cannot be present
in pipitzahoic acid. In hydroxybenzoquinone there are, however,
three benzene hydrogen-atoms which it is possible may be displaced
by hydrocarbon-groups ; in other words, it is conceivable that in
pipitzahoic acid either one hydrogen-atom is displaced by C,H,;, or
that two are displaced by C,H, or three by C,H. The group C,H);
is a radicle of a hydrocarbon of the C,H2,+, series. If two hydrogen-
atoms are displaced by distinct radicles which in sum are equal to
C,H, one of these must be a C,H»; and the other a C,H.» _,
radicle, and in like manner if three radicles are present two must
belong to the C,Hen+:, and one to the C,H», series. In short,
whether one, two, or three of the hydrogen-atoms in the hydroxy-
quinone be displaced by hydrocarbon radicles, one side-chain in
pipitzahoic acid will contain a pair of doubly-linked carbon-atoms.”

Hence arises the question, how many and what side-chains does
the monohydroxybenzoquinone present in pipitzahoic acid contain ?
It may be fitting at once to mention here that we have so far only
succeeded in answering the first part of the question, and in showing
by means of an indirect method that pipitzahoic acid most probably
contains two side-chains.

Before proceeding to the description of our investigations, we must
point out that several months after the publication of the paper
which one of us read on pipitzahoic acid, on August 4, 1884, at
a meeting of the Niederrheinische Gesellschaft fiir Natur- und
Heilkunde in Bonn, a paper was published by Mylius (Ber., 18,

3 B2

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

480) on the same subject, and a short time afterwards a second com-
munication (ibid., 18, 936) followed by the same author. We shall
take the opportunity at different points in this paper of comparing our
results with those obtained by Mylius. We do not agree with this
author’s proposal to give pipitzahoic acid the name perezone, for it
appears to us of little purpose to exchange the already well-known
name pipitzahoic acid (pipitzahoinsaeure) for a name which gives us
just as little idea of the constitution of the substance as the old one.
When the structure of pipitzahoic acid has been fully made out, a more
rational name will suggest itself without any difficulty.

1. Preparation of Pipitzahoic Acid.

For 50 grams of nearly pure pipitzahoic acid, one of us is indebted to
Mr. Vigener; we obtained the remainder of that which we used—some-
what over 200 grams—by extraction from 6 kilos. of Radix perezie.
The roots were pulverised as finely as possible and boiled for half
an hour with 10 times their weight of alcohol on the water-bath.
The solution poured off from the insoluble residue was filtered
and about nine-tenths of the alcohol distilled off. On pouring
this concentrated solution into about five or six times its volume
of water at 50° C., a golden-yellow, crystalline precipitate of
pipitzahoic acid was produced. The filtrate from this precipitate
is light brown coloured; in thin layers it is nearly transparent,
but no precipitate is discernible even with the help of a pocket
lens. By extracting with ether and evaporating the solution, or by
evaporating the filtrate itself under the air-pump or on the water-bath,
considerable quantities of a dark brown, tarry product are obtained
which we did not submit to any further examination.

The residue after extraction with alcohol still contains a consider-
able quantity of pipitzahoic acid, and must be again treated twice in
the above-mentioned way; in our experiments the third extract
contained only very small quantities of the acid.

Three portions of roots were treated with the following results :—

200 grams roots gave 10°8 grams pipitzahoic acid = 3°6 per cent.
1600 grams - 55°7 grams - = 3'5 m
4000 grams » 147°0 grams ~ = 37 “

The dried crystalline precipitate melted at 100°, whereas pure
pipitzahoic acid melts at 103—104°. Mylius gives 106—107°. The
pipitzahoic acid was not further purified for the preparation of the
various derivatives.

Pure pipitzahoic acid may be easily obtained by crystallisation from
dilute alcohol or benzene, and then consists of beautiful, golden-yellow,

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID,

flat plates. It is rather difficultly combustible; burnt with lead
chromate—

(1.) 0°1850 gave 0°4870 CO, and 0°1384 H,0.

On combustion with lead chromate—
(2.) 0°1773 gave 0°4673 CO, and 0°1329 H,0.

Burnt with copper oxide in a current of oxygen—

(3.) 0°2184 gave 0°5760 CO, and 0°1588 H,0.
(4.) 02102 ,, 0°5549 0°1500 ,,
(5.) 01963 ,, 05181 01443 ,,
(6.) 01960 ,, 0°5208 01456 ,,
(7.) 0°1421 ,, 03768 01048 _,,

Found.

Calculated for -— ~
C,H 903. 5 II. III. IV. V. Va VI.

C.... 72°58 71:79 71°88 71°94 72°00 72°03 72°47 72°37
H... 806 831 833 808 %793 816 826 819

Pipitzahoic acid sublimes easily; it distils with steam, but it is

decomposed by protracted boiling with water. It dissolves readily in
alcohol, ether, chloroform, glacial acetic acid and benzene; it is also
somewhat soluble in light petroleum. Hot glacial acetic acid appears
to decompose it.

2. Salts of Pipitzahoic Acid.

Weld has examined several salts of pipitzahoic acid. He has also
drawn attention to the characteristic colour of their solutions, which is
like that of potassium permanganate. He found that the barium,
calcium, sodium and lead salts are decomposed by carbonic anhydride.
Weld also analysed the copper and silver salts. We endeavoured to
prepare the potassium and silver salts in order to obtain ethers of
pipitzahoic acid: we could only obtain the potassium salt in the
form of a black tarry mass, but were more successful in preparing the
silver salt by adding silver nitrate in excess to an aqueous solution
of the potassium salt, whereby the violet-brown precipitate which
Weld describes was formed. But this precipitate was very difficult to
filter, remaining in a frothy condition, even after shaking, and is there-
fore very difficult to obtain in a pure state. After washing with
water, in which it is insoluble, and drying in a vacuum desiccator, the
silver salt forms a purplish-blue, amorphous powder, which is easily
soluble in alcohol, very slightly soluble in water, and insoluble in
ether. Analysis gave the following result :—

714 ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

0-2000 silver salt gave 0°0600 silver.
Calculated for
C,;HiO3Ag. Found.
B® veces eeccese 30°00

3. Attempt tv Prepare an Ethyl Ether of Pipitzahoic Acid.

As ethyl iodide gave tarry products with the silver salt we used
ethyl bromide. After heating the silver salt with five times its weight
of ethyl bromide in a closed tube for one hour in the water-bath,
considerable pressure was observed on opening the tube. The liquid
portion, after filtering from silver bromide, was freed from excess of
ethyl bromide by distillation or spontaneous evaporation, and a dark
red liquid then remained from which a very small amount of a crys-
talline substance gradually separated; this substance crystallised
from alcohol in white prisms, melting at 141° C. We did not obtain
sufficient for a combustion.

4. Attempt to prepare an Acetic Ether of Pipitzahoic Acid.

The acid was heated with three times its weight of acetic anhydride
for three days in a sealed tube in the water-bath. The completion
of the reaction is easily determined, for not only does the colour of the
liquid become lighter, but whilst at first on cooling the tube pipitzahoic
acid crystallises out, this is not the case when the reaction nears
completion. On evaporating the contents of the tube over soda-lime,
the product of the reaction separates in large, colourless crystals,
which may be purified by recrystallisation from alcohol. The
compound thus prepared forms well-characterised prisms; it melts at
115° C., and is easily soluble in alcohol, ether, chloroform and glacial
acetic acid.

Dr. Hintze was kind enough to examine the crystals, and communi-
cated the following description to us :—-

“ System: orthorhombic.

Ratio of the axes a:b: ¢: = 0°62933 : 1 : 0°84507
Observed faces: c = (001)0P
p = (110)coP
m = (120)coP2
= (011)Poo
q= (012)3Pco.

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

Angles of the Normals.

Calculated. Found.
— : 40° 12’
22° 54’ 22 48
17 18 17 24
64 22 64 20
19 21 19 23
103. 4 lu2 48
110: O11 — 69 52

120: 012 59 39 60 0

‘*The crystals are of a short prismatic habit ; the vertical faces are
distinctly striated.”

It is noteworthy that the compound is white, whereas the ethers of
the hydroxyquinones and the quinones themselves are generally
coloured. Its behaviour with caustic potash is also remarkable: the
acid ethers of the hydroxyquinones are easily decomposed by this
reagent, the hydroxyquinone being regenerated, whereas the product
of the action of acetic anhydride on pipitzahoic acid appears to remain
unchanged on treatment with caustic potash, for the substance thus
obtained melted at 113°, and the violet-blue colour of the potassium
salt of pipitzahoic acid did not make its appearance. Since the
decomposition of this compound, melting at 115° C., into pipitzahoic
acid and acetic acid did not succeed, we cannot with certainty main-
tain that we have obtained the acetic ether of pipitzahoic acid, not-
withstanding that the results of analysis agree very well with the
formula of a monacetate: thus

(1.) 0°1523 gave 0°3909 CO, and 0°1105 H,0.

(2.) 01024 ,, 02645 , 00700 ,,

(3.) 0°1454 ,, O8739 , 01004 ,,

Found.

001 : 011
001 : 012
011 : 012
110 : 110
110 : 120
120 : 120

Calculated for r .
C,;Hjg03-C,H;0. II. III.
70°35 70°00 70°44 70°13
8:06 7°59 7°67

5. Attempt to reduce Pipitzahoie Acid.

(a.) By Stannous Chloride—A solution of pipitzahoic acid may be
decolorised by the action of sulphurous acid, or zinc-dust and acetic
acid, or stannous chloride and hydrochloric acid, a compound being
formed which is exceedingly soluble in water. We gave stannous
chloride the preference because the tin can be easily and completely
removed by means of sulphuretted hydrogen. An alcoholic solution

716 ANSCHUTZ AND LEATHER, PIPITZAHOIC ACID.
of pipitzahoic acid was poured into a solution of stannous chloride in
hydrochloric acid, and the tin precipitated by sulphuretted hydrogen
gas from the colourless solution heated to about 50°. The sulphides
of tin were filtered off as rapidly as possible, and the clear colourless
solution thus obtained was concentrated in a stream of carbonic anhy-
dride on the water-bath under a pressure of about 40mm. During this
operation we noticed that when the solution became somewhat con-
centrated, almost colourless oily drops commenced separating; that
as the distillation proceeded these became yellow, then brown, and
lastly, when almost all the water had gone over, they remained as a
tar at the bottom of the flask.

Exactly similar results were obtained on using sulphurons acid.

The tar obtained by the decomposition of the reduced pipitzahoic
acid recalls in some respects the tar remaining after the evapo-
ration of the brown filtrate obtained in the preparation of
pipitzahoic acid from Radix pereziw. If air be passed through an
aqueous solution of the reduced pipitzahoic acid, the liquid becomes
yellow, and the quinone is rapidly precipitated. Further attempts to
isolate the hydroquinone, which is apparently very unstable, were given
up, the more so because other experiments removed all doubt of the
quinone-like character of pipitzahoic acid.

(b.) Distillation of Pipitzahoic Acid over Zine-dust.—One gram of
pipitzahoic acid was intimately mixed with 20 grams of zinc-dust,
filled into a hard glass combustion-tube, another 20 grams of zine-dust
filled in in front, and distilled in a current of hydrogen at a low red
heat. Twenty such operations produced 5 grams of a dark red oil,
which, on distillation, passed over at from 50° to 330° C., no portion
exhibiting a more constant boiling point than another. As much
larger quantities of the acid must have been used to secure a definite
result, we have for the present postponed the repetition of this experi-
ment.

6. Action of Sulphuric Acid on Pipitzahoic Acid.

On treatment either with concentrated or with dilute {1 vol. : 1 vol.)
sulphuric acid, pipitzahoic acid appears to char-at once. No more
satisfactory result was obtained when the sulphuric acid was diluted
with acetic acid. On the other hand, hydroxypipitzahoic acid, which
is described later on, reacts very easily with sulphuric acid on warm-
ing, as Mylius found, whereby a yellow substance, named by him
“perezinone,” which contains the elements of one molecule of
water less than hydroxypipitzahoic acid, is formed. We have also
made the experiment, and can corroborate Mylius’s results.

ANSCHUTZ AND LEATHER; PIPITZAHOIC ACID.

Action of Bases on Pipitzahoic Acid.

During the past few years various authors, especially Zincke,
V. Meyer, and Goldschmidt, have shown that the quinones behave in
a characteristic manner when submitted to the action of amines, of
phenylhydrazine, and of hydroxylamine. In order to prove, there-
fore, the quinone-like nature of pipitzahoic acid, we have subjected
it to the action of these bases, and have found that it very readily
reacts with the amines, in the same way as do the quinones of the
benzene series.

1. Amine-derivatives of Pipitzahoic Acid.

The investigations of Zincke and several of his pupils have shown
that the benzoquinones react with primary and secondary amines
generally in such a way that one molecule of the quinone withdraws
from another molecule one atom of hydrogen, and a second atom of
hydrogen from the amine. The one molecule of the quinone thus
becomes reduced, and the residue of the second molecule combines
with the amine-group. This reaction may be repeated with the
alkylated or phenylated monamidoquinone, and an alkylated or
phenylated diamidoquinone formed. Sometimes this reaction is
accompanied by another chemical change, which consists in one of the
quinone oxygen-atoms being displaced by an amido-group, the oxygen
combining with the two hydrogen-atoms of the amine to form a
molecule of water; this is the case, for instance, in the formation of
dianilidobenzoquinone anilide.

The compounds formed by these reactions are generally solid
crystalline substances of a dark blue, violet-blue, or red colour. They
are prepared by adding a primary or secondary amine to a solution of
a quinone in alcohol or glacial acetic acid. Even in the cold the
liquid becomes dark coloured, and if the solution be not too dilute,
the newly-formed substance crystallises out after a short time. If
these substituted amidoquinones are boiled with dilute sulphuric acid
or hydrochloric acid in alcoholic solution, each of the amine-groups is
displaced by hydroxyl, the amine being reproduced. This last-
mentioned reection affords an excellent method for the preparation of
the hydroxyqninones, and offers at the same time a possibility of
determining the number of the amine-groups which have been intro-
duced. On account of these facts we investigated the action of
aniline, orthotoluidine and paratoluidine on pipitzahoic acid.

O
(a.) Anilidopipitzahoic Acid, CisHis {0 .—To prepare this
NH-C,H; ;

compound excess of aniline is added to an alcoholic or acetic acid

718 ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

solution of pipitzahoic acid. At first the liquid turns violet-red,
and if sufficiently concentrated becomes thick on standing, owing to
the separation of crystals, which consist of small violet prisms. If
these crystals are filtered off and the mother-liquor be allowed to stand
exposed to the air, the hydroquinone of pipitzahoic acid formed in the
reaction becomes oxidised, and the pipitzahoic acid thus regenerated
reacts in the same way with the excess of aniline. We thus obtained
125 grams of anilidopipitzahoic acid from 10 grams of pipitzahoic acid,
whereas, according to theory, we ought to have obtained 13°6 grams.
The formation of the anilide takes place in accordance with the
following equation :—
O

2
2CisHio(O2)-OH + NHyC.Hs = CsHisy OH + CjsHw(OH)s.
NH-C,H;

Anilidopipitzahoic acid forms dark purple, shining needles, which
melt at 133° C. In alcvhol, ether, benzene, chloroform, and acetic
acid, it is easily soluble, and may be crystallised from warm alcohol
or acetic acid. Analysis gave the following results :—

(1.) 0°3097, burnt with powdered CuO, gave 0°8406 CO, and

0°2181 H,0.
(2.) 02524, burnt with powdered CuO, gave 0°6829 CO, and
0°1756 H,0.
(3.) 0°2277, burnt in open tube with oxygen, gave 0°6178 CO, and
0°1581 H,0.
(4.) 0°1974, burnt in open tube with oxygen, gave 0°5356 CO, and
0°1362 H,0.
(5.) 0°4952 gave 18°6 c.c. N at 25° C., under 7631 mm.
Found.
Calculated for r A \
C2, Ha,U5N. i. ll. UL Iv. VY.
© secees 74°34 7403 73°77 7400 7400 —
H weeeee = 7°37 78l 772 #771 76606 —
N wcccee 4:13 — — — — 418
O;
(b.) Orthotoluidopipitzahoic Acid, Cubtad Ob , is
NH{[2)C,H,[1]CH,

prepared from orthotoluidine and pipitzahoic acid in the same way as
anilidopipitzahoic acid, although a good yield may more easily be
obtained by making use of an alcoholic solution. 5 grams of pipitzahoic
acid gave 5 grams of toluide instead of 7:1 grams. Orthotoluido-
pipitzahoic acid crystallises from warm alcohol in needles which melt
at 109—111° C. Mylius observed the melting point 135—136°C. How
this difference is caused we are unable to state. Orthotoluido-

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID. 719

pipitzahoic acid does not show any metallic lustre; it is easily soluble
in alcohol, ether, benzene, chloroform and acetic acid. Analysis gave
the following results :—

(1.) 0°2059 substance gave 0°5620 CO, and 0°1475 H,0.

(2.) 0°4587 - » 167 cc. N. at 11° C. and 752 mm.

Found.
Calculated for

0;
(c.) Paratoluidopipitzahoic Acid, Cublnd Oi » is
NH[4]C,H,[1]CH;
prepared in the same manner as anilido- and orthotoluido-pipitzahoic
acid. This paratoluide, after having been several times recrystallised
from alcohol, forms long, blue needles, without metallic lustre, melting
at 136° C. It is easily soluble in ether, alcohol, benzene, chloroform
and acetic acid. Analysis gave the following results :—

(1.) 0°2478 substance, burnt with PbCrO,, gave 0°6689 CO, and

0°1809 H,0.

(2.) 0°2187 substance, burnt with PbCrO,, gave 0°5914 CO, and
01589 H,0.

(3.) 0°2341, burnt in a current of oxygen, gave 0°6417 CO, and
0°1688 H,0.

(4.) 0°4416 gave 15°8 c.c. N at 19° C., and 756 mm.

Found.

Calculated for - ~y
C2H2,03N. I. II. III. V.

74°79 73°62 73°75 74°75 —
7°65 8:11 8:07 8°01 —
— _ — 4°14

Anilido-, orthotoluido- and paratoluido-pipitzahoic acids behave
with reagents as follows:—On treatment with stannous chloride dis-
solved in hydrochloric acid, or zinc-dust and acetic acid, their solutions
are decolorised, but on exposure to the air the original compound is
reproduced, and hence it would appear that the quinone oxygen-atoms
remain intact in these three derivatives. Anilido- and the two toluido-
pipitzahoic acids dissolve in concentrated sulphuric acid with a blue
colour ; if these solutions be poured into a large quantity of water no
precipitates are formed, but simply a purple-red liquid. The three
acids dissolve with difficulty in alkaline solutions. The concentrated
solutions in alkalis are of a violet-red colour, which, on the addition

720 ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

of much water, passes into green. By heating with ammonia in a
closed tube the aniline and toluidine groups become displaced by the
amido-group. It is noteworthy that we have obtained in these three
derivatives no substances which may be looked upon as diamine-deri-
vatives, and from this it may be inferred that there is only one
benzene hydrogen-atom in pipitzahoic acid.
O,
(d.) Amidopipitzahoie Acid, C,sHis {9 .—If the anilide or the
NH,
toluides of pipitzahoic acid are boiled with ammonia they suffer a
change, their violet-blue colour passing into red-brown. In order to
prevent the escape of ammonia these operations were carried out in
sealed tubes. On account, however, of the fact that even after six
hours’ heating at 100° C., considerable quantities of undissolved
substance remained in the tubes, we used alcoholic ammonia and
heated as before for six hours in the water-bath. In each case the
violet-blue colour of the solutions changed to blue.

The tubes in which we had used aqueous ammonia contained blue
liquids, and also insoluble substances, consisting of dark-brown
crystals contaminated with some tar. On concentrating the blue
filtrates large quantities of dark brownish-red, shining crystals sepa-
rated, and the blue colour of the solutions changed to brown. The
brown crystals from the tubes, mechanically separated from the tar,
melted at 135—140° C., the crystals from the mother-liquors at
146—149°, whether the material used had been the anilide or either
of the toluides of pipitzahoic acid. All the products consist of
the same substance in different degrees of purity. We did not,
however, further investigate the action of aqueous ammonia, as we
obtained the same compound by the use of alcoholic ammonia, in
which latter case no tar was produced. On evaporation of the alcoholic
solutions, the smell of aniline or ortho- or para-toluidine could easily
be detected. In order to determine the quantity of the amine which
was in each case set free, we subjected to distillation with steam
the three blue solutions obtained by heating with alcoholic ammo-
nia. Shortly after the commencement of the distillation, the blue
colour of the solutions changed to brown, and large quantities of dark
reddish-brown crystals were precipitated which melted at 150° C.,
and consisted of amidopipitzahoic acid.

Four grams of anilide was heated for six hours at 100° C. with
about three times the theoretical quantity of concentrated alcoholic
ammonia. The contents of the tubes was then distilled with steam,
and the distillate containing ammonia and aniline was extracted three
times with ether. This ethereal extract was washed with water until
it no longer turned litmus-paper blue; an excess of hydrochloric acid

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID. 721

was then added, and the solution evaporated to dryness on the water-
bath. The residue consisted of nearly pure aniline hydrochloride, and
weighed 1:2 gram, whereas according to the equation—

O, O,
C,;Hs OH + NH; — CisHi, OH a C.H;N H,,
NHC,H; NH,

we ought to have obtained 1°5 grams of aniline hydrochloride.

Operating in the same way we obtained from 2°1 grams of para-
toluidopipitzahoic acid, 0°7 gram of paratoluidine hydrochloride
instead of 0°85 gram, and 1°45 grams amidopipitzahoic acid instead
of 2°1 grams.

When crystallised from alcohol, amidopipitzahoic acid forms red-
dish-brown, shining crystals which melt at 151° C. It dissolves in
solutions of the alkalis and ammonia with a deep blue colour, which
is not changed on addition of water. On boiling with water the
ammonia salt is decomposed, amidopipitzahoic acid is precipitated,
and the blue colour of the solution disappears. Amidopipitzahoic acid
is insoluble in water, but is easily soluble in alcohol and ether. Sub-
jected to reducing agents such as stannous chloride and hydrochloric
acid it dissolves easily, the solution being decolorised. If these
colourless reduced solutions are exposed to the air, or if air is injected
into them, the hydro-compounds are oxidised and amidopipitzahoic
acid is precipitated. It dissolves in cold concentrated sulphuric acid
forming a reddish-brown solution, which becomes red on addition
of much water. On boiling with dilute sulphuric acid, amido-
pipitzahoic acid is converted into hydroxypipitzahoic acid. Analysis
gave the following results :—

(1.) 0°1493 substance gave 0°3713 CO, and 0°1128 H,0.
(2.) 0°3124 - 16°2 c.c. N at 20° and 753 mm.
(3.) 0°4175 - 19°5 ~ 13° , 72 .
Found.
Calculated for Fon ~
CysH.,NO3. i. II.
68°44 67°83 —-
8°38 -=
— 5°95 5°55

(e.) Proof of the Identity of Mylius’s “ Perezonoxime”’ with Amido-
pipitzahoic Acid.—In order to characterise pipitzahoic acid as a quinone
Mylius treated it with hydroxylamine. According to Goldschmidt
the benzoquinones are only reduced to hydroquinones by the action
of hydroxylamine, whereas by the action of hydroxylamine hydro-
chloride paranitrosophenols are formed (quinone-oximes). [It is still

722 ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.
doubtful whether the nitrosophenols are properly regarded as quinone-
oximse.] The properties of pipitzahoic acid are such that it appears to
belong to the group of benzoquinones; it was, therefore, somewhat
surprising that Mylius obtained a well-characterised compound by the
action of hydroxylamine on pipitzahoic acid, analysis of which gave
results agreeing with the formula C,,H,O;N, and which Mylius
considered as the “ perezonoxime ” (oxime of pipitzahoic acid) formed
according to the reaction—

(CisH»0)O, + NH,-OH = (C\sH»O)ON (OH) + H,0.

Mylius has his doubts as to the correctness of this supposition, on
account of the fact that by the action of boiling hydrochloric acid on
this “perezonoxime,” ammonium chloride and hydroxypipitzahoic acid,
which has already been mentioned, and which will be described at a
later period, were formed. He says, in fact :—‘‘ The action of concen-
trated hydrochloric acid on ‘perezonoxime’ takes place, therefore,
according to the equation—

C,;H,,NO; + HCl + H,O 7 C,s;H»O, + NH,,HCIl.

“Tt is a similar change to that which the amides undergo when, in
consequence of the exchange of the amido-group for the hydroxyl-
group, they become acids. Now, the hydroxypipitzahoic acid is really
a description of acid, and it is easily conceivable that the hydroxyl-
amine-derivative in question is its amide, and must be represented by
the formula C,;;H,O;;NH,. In this case the equation expressing its
formation would be the following—

CisH»0; + NH,OH = C,;H,,O,NH, + H,0.”

Mylius further says, “A more thorough study of this compound is
requisite before a decision in favour of the one or the other view can

be arrived at.”
The discovery that amidopipitzahoic acid is formed by the action of

ammonia on anilido-, and on ortho- or para-toluido-pipitzahoic acid
permits of a decision as to the real nature of Mylius’s “ perezonoxime ”
being easily arrived at; in fact, we have found by a comparison of the
“perezonoxime,” prepared according to Mylius’s directions, with
amido-pipitzahoic acid that the two substances are identical. Both
preparations melted at 151° (according to Mylius “ perezonoxime” melts
at 153—154°). They possess the same colour; they crystallise in
the same forms from alcohol ; they behave in the same manner with
solvents, with alkalis, and with concentrated sulphuric acid; and on
boiling their alcoholic solutions with dilute sulphuric acid they both
give hydroxypipitzahoic acid. We have also analysed Mylius’s “ pere-

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID. 723

zonoxime,” and, as was to be expected, have obtained results which
agree with the formula of amidopipitzahoic acid——

(1.) 0°1780 substance gave 0°4457 CO,, and 0°1391 H,0.
(2.) 0°4742 . 21-6 c.c. N at 18°, and 754 mm.

Found. Found by
Calculated for ——*+—-—, Mylius.
C,;H,NO3. I. Il. —_——
68°44 68:28 — 68°25 om
8°68 _ 811
— 528 at

As there is no longer any doubt that pipitzahoic acid is a quinone,
proof of the identity of “ perezonoxime” with amidopipitzahoic acid
having thus been given, it follows that hydroxylamine may act on the
quinones in a manner hitherto unknown, and may cause the intro-
duction of the amido-group in place of one of the benzene hydrogen-
atoms. The formation of amidopipitzahoic acid from ammonia and
anilido- or ortho- or para-toluidopipitzahoic acid also gives promise of
a method for the preparation of amidoquinones which appears to be
of general application, as indeed is shown by an experiment with
hydroxythymoquinone described later in the paper.

2. Action of Hydroxylamine Hydrochloride on Pipitzahoic Acid.

Having obtained the above-mentioned results with hydroxylamine,
it appeared to us necessary to investigate the action of hydroxylamine
hydrochloride in the hope of obtaining the pure quinoneoxime, which,
if pipitzahoic acid were really a paraquinone, ought to belong to the
class of nitrosophenols. We have, however, not as yet been successful.
Various experiments which were carried out at temperatures ranging
between 80° and 100° gave nothing but tarry products, although the
reaction appears to take place better in the cold. The pipitzahoic
acid was dissolved in about 20 times its weight of 50 per cent. cold
alcohol, and after adding the requisite quantity of hydroxylamine
hydrochloride, as well as a few drops of hydrochloric acid, the solution
was set aside. At the end of two days its colour had become much
darker, and when at the end of eight days the solution was poured
into an excess of water, a red precipitate was formed, which became
partially crystalline on being well shaken up. 5 grams of pipitzahoic
acid yielded 4°9 grams of this precipitate, which consisted of a red,
crystalline substance and some tar. As this tar proved to be about
as soluble as the crystalline substance in alcohol, ether, benzene,
chloroform, glacial acetic acid, carbon bisulphide and light petroleum,
we have not yet been able to prepare the latter in the pure state.

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

3. Action of Phenylhydrazine on Pipitzahoic Acid.

Phenylhydrazine like hydroxylamine acts energetically on the
quinones (Ber., 16, 1563 ; 17, 1809, 3026) ; indeed, phenylhydrazine is
decomposed by paraquinones such as benzo- tolu- and thymo-quinone
with liberation of nitrogen, whereas phenylhydrazine hydrochloride
reacts less energetically, but no well-characterised derivatives have
been obtained from these quinones. On the other hand, according to
Zincke, S-naphthaquinone, phenanthraquinone and other similar
quinone-derivatives react well. Especially interesting is the identity
which Zincke has proved to exist between the a-naphthaquinone-
hydrazid, obtained by the action of a-naphthaquinone on pheny!l-
hydrazine hydrochloride, and benzene-azo-«-naphthol.* As Zincke
says that he has prepared a well-characterised compound from
phenylhydrazine and hydroxythymoquinone, which pipitzahoic acid
so much resembles, we have treated pipitzahoic acid both with
phenylhydrazine and also with phenylhydrazine hydrochloride.

If alcoholic solations of pipitzahoic acid and phenylhydrazine be
mixed, the liquid becomes dark-coloured and gas is evolved. 1 gram
of pipitzahoic acid and 0°5 gram of phenylhydrazine produced 50 c.c.
of nitrogen at 20°. The volume of nitrogen evolved is slightly greater
than that which would represent the half of the total nitrogen in
05 gram of phenylhydrazine, on evaporation, the alcoholic solution
gave nothing but a tarry product. Phenylhydrazine hydrochloride
gave no better result; we added 1 gram of phenylhydrazine hydro-
chloride in 10 c.c. of water to 1 gram of pipitzahoic acid in 100 c.c. of
alcohol, and after allowing the mixture to stand for two days at the
ordinary temperature, we poured the dark-coloured solution into a
large excess of water, whereby a brownish-red, tarry precipitate was

* I should like to take this opportunity of drawing attention to the following
facts. The paraquinones are well known to be powerful oxidising agents. In the
formation of a-naphtholazobenzene from a-naphthaquinone and phenylhydrazine,
an intermolecular oxidation apparently takes place, the hydrazine-group losing one
hydrogen-atom and the “quinone” oxygen-atom being reduced. It appears to me
that the conditions under which the formation of nitrosophenols from the para-
quinones and hydroxylamine takes place are exactly similar. Another fact must
also here be taken into consideration, viz., that hydroxylamine is a reducing agent.
I think agreement with other analogous reactions is better preserved in assuming
that in the formation of the nitrosophenols from the quinones and hydroxylamine,
through an intermolecular reducing process the “ quinone’’ oxygen-atom changes
to hydroxyl, than that in the formation of nitrosophenol from nitroso-dimethyl-
aniline the “ nitroso”-group changes to the “oxime’’-group. It is still more
difficult to explain the latter reaction, using the “‘ oxime” formule for the para-
nitrosophenols, than the transformation of the nitrosophenols into the nitrophenols.
The orthoquinones stand in close relationship to the diketones, and what may be
proved for them does not necessarily apply to the paraquinones.—ANscHiTz.

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID. 725

thrown down which appeared to be partially crystalline, but possessed
no characteristics which invited an experimental investigation. In
this experiment nitrogen was not evolved. Both with phenylhydra-
zine and its hydrochloride, pipitzahoic acid appears to behave in an
exactly similar manner to benzo- tolu- and thymo-quinones.

APPENDIX.
Action of Ammonia on Anilidohydroaythymoquinone.

The exclusive formation of monamido-compounds by the action of
aniline, ortho- and para-toluidine on pipitzahoic acid suggested the
hypothesis that there might only be one benzene hydrogen-atom in
pipitzahoic acid. As it appeared to us advantageous to compare the
behaviour and properties of a known hydroxyquinone, in which also
there was only one benzene hydrogen-atom, with that of pipitzahoic
acid, we chose hydroxythymoquinone, as it has been the best
investigated of the well-known similar quinones, and also can be
obtained with comparative ease. Like pipitzahoic acid, hydroxy-
thymoquinone forms with aniline and paratoluidine only a mon-
anilido- or paratoluido-derivative. By boiling with dilute sulphuric
acid the anilidohydroxythymoquinone is changed to dihydroxythymo-
quinone. So far we have only examined the action of alcoholic
ammonia on anilidohydroxythymoquinone in order to thus generalise
the reaction by which amidopipitzahoic acid is formed from anilido-
pipitzahoic acid.

Following the directions of H. Schulz we prepared from thymo-
quinone (Ber., 16, 898) first dimethylamidothymoquinone, and from
this monhydroxythymoquinone, the melting point of which we
found to be 164—165° (Schulz, 166°). Hydroxythymoquinone crystal-
lises in yellow plates which closely resemble pipitzahoic acid. The
requisite quantity of aniline was added to an alcoholic solution of
monhydroxythymoquinone, and by this means we obtained the anilido-
hydroxythymoquinone described by H. Schulz. This compound
crystallises in violet-black needles, which melt at 184—135°, and may
easily be purified by recrystallisation from alcohol.

CH;

Diamidothymoquinone, Ce a. , or Amidohydroxythymoquinone-
2

(NH).
CH;
C;H,

imide, OO, Ray gram of anilidohydroxythymoquinone was

OH
\NH
. XLIX. 3 ¢

726 ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

sealed up in a tube with 20 c.c. of the strongest alcoholic ammonia
and heated in the water-bath; the tubes being now and then well
shaken. After three hours’ heating it was noticed that the colour of
the solution had changed from red-brown to deep blue. At the end
of two days a large quantity of dark-blue, shining crystals separated.
In order to estimate the amount of separated aniline, we operated
exactly as before in the case of amidopipitzahoic acid and in three
experiments obtained the following results :—

(1.) 1 gram of the anilide with 15 c.c. of alcoholic ammonia gave
0°45 gram of the crystalline product and 0°95 gram of aniline
hydrochloride, m. p. 193°.

(2.) 4 grams of the anilide in 60 c.c. of alcoholic ammonia gave
2°1 grams of the crystalline product and 0°95 gram of aniline
hydrochloride.

(3.) 4 grams of the anilide in 40 c.c. of alcoholic ammonia gave
2-0 grams of the crystalline product and 1 gram of aniline

hydrochloride.

We shall refer to these figures again.
The dark-blue crystals show a marked metallic lustre and give a

green powder on pulverisation. They do not melt but sublime at a
high temperature. In water, ether, benzene, chloroform, carbon
bisulphide and light petroleum this compound is insoluble, and in hot
alcohol it dissolves only to a very limited extent to a blue solution.
Boiling glacial acetic acid on the other hand dissolves it in consider-
able quantities, and on cooling the solution nearly the whole crystal-
lises out again in the form of black, shining plates. It dissolves
very easily even in dilute hydrochloric acid, which is indicative of the
presence of an amido-compound. Analysis of the crude product
gave figures which agree with the formula of the hitherto unknown

diamidothymoquinone.

(1.) 0°2031 gave 0°4563 gram CO, and 0°1398 H,0.
(2.) 04384 , 545 cc. nitrogen at 7° and 743 mm.
Found.
Caleulated for pee
peenenias 61°85

Singularly enough, we obtained different results on analysing the
substance crystallised from hot acetic acid and freed from the solvent
by being placed in a vacuum of 10 mm. over soda-lime; these results

agree, however, closely with one another.

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.
(1.) 0°2149 gave 0°4630 CO, and 0°1450 H,0.
(2.) 01857 ,, 04000 ,, O1193 ,,

The substance was now again recrystallised from acetic acid.

(3.) 0°1788 gave 0°3876 CO, and 0:1166 H,0.
(4.) 0°2196 ,, 23 cc. N at 9°5° and 753 mm.

Found.

Calculated for Calculated for r
CyoH}4N2O. CipH,,N2O, + $(C2H,0,). ae II. III.
61°85 58°92 58°76 58°83 59°12
7°21 7°14 749 714 5°24

14°43 12°50 —- —- — 12°58

Of the substance used in analyses 3 and 4 a portion, after standing
another two days over soda-lime in a vacuum, was pulverised in
an agate mortar (when the smell of acetic acid was distinctly per-
ceptible), and heated in a Liebig’s drying tube at 100° (Liebig’s
Annalen, 228, 303) under a pressure of 30 mm. in a slow current of
air until the weight became constant—

(1.) 0°5756 lost 0°0768 of its weight.
(2.) 0°3676 ,, 0°0492.
Found.

Calculated for oa
C,H, ,N,0, + $(C,H,0,). I. II.
C.H,0, 13°34 13°44

The substance thus freed from acetic acid was again analysed :—

(1.) 0°1956 gave 0°4412 CO, and 0°1269 H,0.
(2.) 0°2457 ,, 293 c.c. Nat 2° and 766 mm.

Found,
Calculated for

These figures afford proof that the substance is either diamido-
thymoquinone or the isomeric amidohydroxythymoquinone-imide :—

CH;
C;H,
O

Os NHS?
OH
NH,

and that it crystallises from acetic acid with } mol. of the solvent.
3c 2

728 ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

Ammonia acted therefore in a somewhat different manner on anilido-
hydroxythymoquinone to what it did on anilidopipitzahoic acid ; the
aniline-group was in this case displaced by the amido-group, and thus
far the action of ammonia on these aniline-derivatives was the same.
Whilst, however, in the case of anilidopipitzahoic acid, the action of
the ammonia was restricted to the production of amidopipitzahvic
acid, the anilidohydroxythymoquinone was still further altered.
Now whether the second nitrogen-atom displaced the quinone
oxygen-atom in the form of the imido-group or the hydroxy-group in
the form of the amido-group, we must leave for the present unde-
cided. The properties of the product, however, are more those of a
diamido-compound, and we shall therefore speak of it as diamido-
thymoquinone.

The yield of diamidothymoquinone is by no means quantitative, a
dark brown, amorphous substance, easily soluble in alcohol, being also
produced, and this we did not examine further. The following
tabulation of the amounts of diamidothymoquinone and of aniline
hydrochloride obtained from the above-mentioned operations show
that, although the results of experiments No. 2 and 3 are materially
better than that of No. 1, the yield only amounts to five-sevenths of
that theoretically possible.

Found.
Calculated for lam A ~
C,9H,,03:NH-C,Hs. I. II. IIL.
C,oH,N.0; eovece 71:5 45°0 52°0 50°0
C;sH;NH2HCl.... 47°7 15°0 23°8 25°0

Diamidothymoquinone is a feeble base, and dissolves in hydrochloric
acid with a red colour, but by evaporating the solution on the water-
bath the salt is decomposed. On boiling it with concentrated caustic
potash it is dissolved, ammonia being given off, and on acidulating
with hydrochloric acid a red precipitate is thrown down. Stannous
chloride in hydrochloric acid reduces it at once; it being dissolved
to a colourless solution. Agitation of this solution with ether, and
subsequent evaporation of the ether showed that nothing had been
extracted, but if the tin is precipitated with sulphuretted hydrogen
and the sulphides of tin filtered off, the clear colourless solution is
very rapidly oxidised on exposure to the atmosphere and the solution
again turns red, the hydroquinone being oxidised to quinone.

Hydroxypipitzahoic Acid.

In his investigation of the compounds of the quinones with amines
Zincke found that on boiling the alcoholic solutions of these deriva-
tives with dilute sulphuric or hydrochloric acid, the amido-groups are
displaced by hydroxyl.

ANSCHUTZ AND LEATHER: PIPiTZAHOIC ACID. 729

C.H, { . + H,0 = O,H, { % + NR'R"H.

Tn this equation R’ and R” may represent either hydrogen-atoms or
monad hydrocarbon-groups.

Tf an alcoholic solution of anilidopipitzahoic acid be boiled with a
little dilute sulphuric acid, the colour of the solution changes very
rapidly to orange. On addition of water, an apparently homogeneous
red precipitate is formed which may be easily extracted with ether,
and crystallises very beautifully from this solvent. This reaction
takes place just as easily with the two toluides of pipitzahoic acid,
and also, though somewhat more slowly, in the case of amido-
pipitzahoic acid. The substance thus produced was recrystallised
from alcohol, and was obtained in red crystals melting at 127—129°.
Analysis gave the following results :—

(1.) 0°1920 gave 0°4726 CO, and 0‘1414 H,0.
(2.) 0°1652 gave 0°4558 CO, and 0°1350 H,0.

Found
Calculated for —_———
Cys H904. kL. II.
FF ssessens 68°18 67°13 67°12
| Se 7°57 8:18 8:10

The results of both analyses are accordant, but do not agree with
the formula of hydroxypipitzahoic acid, C\jsHy»O,. In preparing this
substance a second time we used benzene as the solvent, and in dissolv-
ing the crude product we found that the hydroxypipitzahoic acid was
mixed with a pale yellow substance, very insoluble in benzene, which
being very easily soluble in alcohol we had not previously observed.
The substance after having been purified by repeated recrystallisation
from benzene, melted at 124°, but the vesults of the analysis were still
less in accordance with the formula C,sH2»0,.

(1.) 0°2261 gave 0°5512 CO, and 0°1656 H,0.
(2.) 0°2319 when burned with PbCrO, gave 0°5600 CO, and
0°1709 H,0.

The substance used for analysis 3 was now again crystallised
from benzene.

(3.) 0°1515 gave 0°3687 CO, and 0°1119 H,0.

Found.
Calculated for r A ~
C,;H90,. a II. III.
FP cnvece 68°18 66°47 66°00 66°37

|: eer 7°57 8°13 8:18 8:20

730 ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

We now altered our method of preparing the acid in so far that we
used the least possible quantity of sulphuric acid. The anilido-
pipitzahoic acid was dissolved in dilute alcohol, and while the solution
was boiling in a retort connected with an inverted condenser, dilute
sulphuric acid was slowly added drop by drop until the blue colour of
the solution changed to red. The product of the reaction separated
almost completely as the solution cooled, and after drying and
recrystallising from benzene melted at 138°, but in other respects it
had the same appearance as the two other specimens melting at 124°
and 129° respectively. By this method of preparation we did not
obtain the least particle of the yellow substance above referred to; it
is also noteworthy that we required far less sulphuric acid when working
in this manner than in preparing the two former specimens. The
analysis of the specimen melting at 138° gave the following results :—

(1.) 0°1579 gave 0°3922 CO, and 0°1121 H,0.
The substance was now recrystallised from benzene.

(2.) 0°1447 gave 0°3588 CO, and 0°1010 H,0.

Found.
Calculated for oa
Ci5H 9094. a II.
ee 68°18 67°74 67°63
ie akenwuws 7°57 7°88 7°75

According to these figures the product must have been almost pure
hydroxypipitzahoic acid, the melting point of which, according to
Mylius, is 133—134°. In one experiment with 3°3 grams of anilido-
pipitzahoic acid we obtained 0°9 gram aniline hydrochloride, instead
of 1°26 gram. The reaction may be represented by the following

equation :—
C,;H,,0,;-NH-C,H; + H,O + HCl = C,sH»O, + C,H;NH,Cl.

In another experiment we used dilute acetic acid as the solvent for
the anilidopipitzahoic acid instead of dilute alcohol, 3 grams being
dissolved in 40 c.c. of acetic acid and 40 c.c. of water; 10 e.c. of
sulphuric acid, composed of 5 c.c. of concentrated sulphuric acid and
5 c.c. of water, was required in order to complete the reaction. But
dilate alcohol is preferable, for the product after crystallisation from
benzene melted at 129—130°, and a comparatively large quantity of
the above-mentioned yellow substance was formed. The decomposi-
tion was complete, however, for we obtained 1:1 gram aniline hydro-
chloride instead of 1:11 gram.

There was not enough of the yellow substance formed in the pre-
paration of hydroxypipitzahoic acid for a careful examination of its
properties. It is certainly not the yellow substance melting at 143°,

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID. 731

mentioned by Mylius, which was prepared by heating hydroxypipitza-
hoic acid with concentrated sulphuric acid, and named by him
“‘ perezinone.” The yellow substance observed by us does not melt, and
is only decomposed on a high temperature being reached ; it is easily
soluble in alcohol and acetic acid, but is almost insoluble in ether and
especially in benzene.

Hydroxypipitzahoic acid dissolves readily in ether, chloroform,
glacial acetic acid, benzene and warm alcohol, and is sparingly soluble
in carbon bisulphide and light petroleum. It crystallises best from
benzene. In solutions of the alkalis the hydroxypipitzahoic acid
forms salts, the aqueous or alcoholic solutions of which are of a purple
colour. With solution of barium hydroxide, it forms a green precipi-
tate which is easily soluble in alcohol. We were not able to obtain a
well-characterised product by heating it with excess of acetic acid
anhydride in a sealed tube at 100°. Hydroxypipitzahoic acid does not
react with aniline; its alcoholic solutions are decolorised by reduc-
ing agents, but we were unsuccessful in all our attempts to isolate
the product of reduction.

Action of Bromine on Pipitzahoic Acid and Hydroxypipitzahoic Acid.

The opinions on the constitution of pipitzahoic acid, founded on
Weld’s experiments, which we expressed in the commencement of this
paper, led us to accept the fact that whether we assume that one, two, or
three of the benzene hydrogen-atoms in pipitzahoic acid are displaced
by hydrocarbon-groups, one of these must contain a doubly-linked pair
of carbon-atoms. In order to test experimentally the truth of this
hypothesis, we very gradually mixed molecular quantities of pipit-
zahoic acid and bromine, both dissolved in chloroform and cooled with
a mixture of ice and salt, and observed that hydrogen bromide was
not evolved on the chloroform being allowed to evaporate, even after
24 hours’ stauding.

Dibromide of Pipitzahoic Acid—The chloroform solution of
pipitzahoic acid becomes dark coloured after treatment with bromine,
and the residue after evaporation of the chloroform is also dark brown-
coloured,

(1.) 1 gram of pipitzahoic acid treated with 0°65 gram of bromine

gave 1°6 grams of the dry product.
(2.) 5 grams of pipitzahoic acid treated with 3°20 grams of bromine

gave 8°20 grams of the dry product.

On estimating the bromine in this dry product by Carius’s method
the following result was obtained :—

0'2059 gave 0:2042 AgBr, equal to 0°0871 Br.

732 ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

Calculated for C,;H 90;Bre. Found.
42°30

The brown-coloured, crude additive product was now triturated with
cold alcohol, whereby we obtained a canary-yellow powder which
when dry melted between 107° and 109°.

02183 gave 01918 AgBr, equal to 0°0817 Br.

Calculated for C};H,0;Bry. Found.
37°42

Experiments made with a view of purifying by recrystallisation
this additive product, which is readily soluble in almost all solvents,
showed the ease with which it is decomposed with separation of
hydrogen bromide.

Dibromide of Hydrowypipitzahoic Acid.—This additive product was
prepared in the same way as the dibromide of pipitzahoic acid, and it
is just as unstable as the latter. On evaporation of its chloroform
solution at the ordinary temperature, it remains as a red, amorphous
powder.

1 gram of hydroxypipitzahoic acid and 0°6 gram of bromine gave
15 grams of the dry additive product.

5°05 grams of hydroxypipitzahoic acid and 3 grams of bromine gave
8 grams of the dry additive product.

These figures prove that the bromine is readily taken up. An
estimation of the bromine in the dry crude product gave the following

result :—

02160 gram substance gave 0°1941 gram AgBr, equal to 0:0826

Calculated for C,;H»0,-Bry. Found.
38°29

The dibromide of hydroxypipitzahoic acid behaves with solvents
very similarly to the dibromide of pipitzahoic acid. It is easily
soluble in alcohol, ether, chloroform and glacial acetic acid, but less
easily in light petroleum and carbon bisulphide ; we could not, how-
ever, recrystallise it from any of these solvents. It melted with
decomposition between 140° and 146°. We endeavoured to obtain it
pure by fractional crystallisation, and the following analytical results
are of fractions obtained from carbon bisulphide (Nos. 1 and 2) and
dilute alcohol (Nos. 3, 4,5 and 6).

ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

(1.) 0°2048 gave 0°1570 AgBr, equal to 0:0670 Br.
(2.) 01981 ,, 01536 0:0655 ,,
(3.) 0°1726 01271 0°0542 ,,
(4.) 0°1507 0°1018 0:0434 ,,
(5.) 0°1994 0°1280 0°0545 ,,
(6.) 0°0977 0°0526 0°0226
Found.

ace

Calculated for _ ~
C);H90,- Bro. ) a II. III. IV. We VI.

37°73 32°71 33°06 31:40 2880 27:32 23:13

The instability of the bromide is clearly proved by these figures :
each successive fraction from alcohol contains less bromine, the
melting point sinking at the same time (the substance used for
analyses 1, 2, 3, 4 melted between 140° and 144°, the substance
No. 5 melted at 135°, and that following (No. 6) at 127°).

Summary.

The acid properties of pipitzahoic acid were recognised by the
discoverer of this substance, Rio dela Loza. The formula deduced by
Weld from his analyses of the acid, and also from the analyses of
several of its salts, is confirmed by our researches. It may be inferred
from Weld’s experiments that pipitzahoic acid is a monobasic acid,
although at the same time it is evident that it does not belong to the
group of carboxylic acids but to that of the phenols, as its alkali metal
salts are decomposed by carbonic anhydride. The beautiful golden-
yellow colour, in addition to the fact that pipitzahoic acid contains
only carbon, hydrogen, and oxygen, suggested that it might be a
quinone; this we have proved to be the case by our experiments.
Hence the functions of the three oxygen-atoms are established.

Discussing the constitution of pipitzahoic acid on the basis of the
above-mentioned facts, it appeared probable that it belonged to the
group of hydroxybenzoquinones, and that it might contain one, two or
three side-chains. The action of amines was therefore studied in the
hope of obtaining evidence of the number of benzene hydrogen-atoms
still present. Our experiments show that pipitzahoic acid reacts with
aniline,and ortho- and para-toluidine with extreme readiness, exactly as
the better-known quinones do, and that always only a single hydrogen-
atom becomes displaced. It may with justice be maintained that a
hydroxybenzoquinone which with aniline, even when this is present in
large excess, gives only a mon-anilide, contains also only one benzene
hydrogen-atom. It may be further suggested that the dihydroxy-
quinone, obtained by the action of acids on this monanilido-oxybenzo-

734 ANSCHUTZ AND LEATHER: PIPITZAHOIC ACID.

quinone, cannot have another hydrogen-atom displaceable by the
anilido-group: hydroxypipitzahoic acid does not, in fact, react with
aniline. Hence it follows that most probably there are two hydro-
carbon side-chains present in pipitzahoic acid. As to the nature of
these side-chains, it has been pointed out that, however many may be
present in pipitzahoic acid, there must be at all events one side-chain

which contains a doubly-linked pair of carbon-atoms: proof of the

truth of this conclusion is given by the behaviour of pipitzahoic acid
and hydroxypipitzahoic acid with bromine. Pipitzahoic acid contains
therefore most probably two side-chains of which the one is a hydro-
carbon radicle of the formula C,Hon+:, the other a hydrocarbon
radicle of the formula C,H2,_,. We are unable at present to express
any opinion as to the nature and position of these radicles.

In order to prove the truth of our ideas of the nature of pipitzahoic
acid by a comparison of its properties with those of a hydroxybenzo-
quinone of known constitution containing two hydrocarbon side-
chains, we have included the best investigated hydroxybenzoquinone,
hydroxythymoquinone, in our experiments. Crystallised from dilute
alcohol, pipitzahoic acid is in colour and crystalline form so like
hydroxythymoquinone crystallised from the same solvent, that the two
are very difficult to distinguish with the unaided eye. Again dihydroxy-
thymoquinone, like hydroxypipitzahoic acid, is a red substance. With
reducing agents pipitzahoic acid and hydroxythymoquinone behave
very similarly, both being converted into colourless hydroquinones,
easily soluble in water, and which on account of their instability have
not yet been isolated. We have not yet succeeded in preparing ethers
of hydroxythymoquinone. On the other hand, we have obtained from
pipitzahoic acid by the action of acetic anhydride, a colourless
compound in good measurable crystals, which according to the
analysis might be an acetic ether. Further we have succeeded in
obtaining by the action of ethyl bromide on the silver salt, a colourless
well-crystallised substance, unfortunately in too small quantity for an
analysis. We are, however, inclined to doubt whether these deriva-
tives really are ethers of pipitzahoic acid: firstly, because we have
not succeeded in regenerating pipitzahoic acid from them; and
secondly, because they are colourless, whereas they ought to be yellow
or orange if the quinone oxygen-atoms are still present. Hydroxy-
thymoquinone and pipitzahoic acid form with aniline and para-
toluidine only a mon-anilide or monoparatoluide. Hydroxythymo-
quinone is prepared in the same way from anilidobydroxythymo-
quinone as hydroxypipitzahoic acid from the anilide of pipitzahoic
acid: treated with alcoholic ammonia, the aniline-group is displaced
in the anilides of both these hydroxyquinones by the amido-group ;
but in the case of the anilidohydroxythymoquinone the reaction goes

BAYLEY: ANALYSIS OF ALLOYS AND MINERALS, 735

further, one of the oxygen-atoms, probably that of the hydroxyl-
group, being displaced by amidogen.

All these facts tend to show that a great analogy exists in the
behaviour of pipitzahoic acid and of hydroxythymoquinone, and
suggest accordingly that the constitution of these substances is
analogous.

LXVIII.—On the Analysis of Alloys and Minerals containing the
Heavy Metals, Selenium, Tellurium, §c.

By Tuomas Bay ey.

Tue metals and metalloids (semi-metals) which can be precipitated as
sulphides from aqueous liquids are naturally divided into two groups.
On the one hand there are those elements forming sulphides insoluble
in alkaline sulphides, and on the other the elements whose sulphides
are soluble. ,

Methods have been devised for separating the one group from the
other by taking advantage of these differences; such methods have
uot been extensively employed in quantitative analysis, although they
have proved useful in the qualitative testing of mixtures. This partial
failure is due to the difficulty of removing the sulphides of the soluble
groups when once they have been precipitated in admixture with the
other sulphides. Repeated digestions with the alkaline sulphide are
necessary if anything approaching to a complete separation is desired.

It is possible, however, by a modification of this method which I
have used for some time, to obtain good results in the analysis of
samples of bronze, gun-metal, fusible and other alloys. Instead of
precipitating both classes of sulphides together and subsequently
treating with the solvents, it is better to altogether prevent the preci-
pitation of the elements forming sulpho-salts. Experience has shown
that the process conducted on this principle is at once accurate and
very easy of execution. The method may be called the sulphide of
sodium method, because in general this is the most convenient solvent
to employ.

About 2°5 grams of the sample is dissolved in the smallest practic-
able quantity of nitric acid of 1°2 sp. gr., or aqua regia may be used
when more convenient, as for instance when tin is present, or when
an estimation of the silica in a mineral is desired.

Ores containing galena are best dissolved by hydrochloric acid and

736 BAYLEY: ANALYSIS OF ALLOYS AND MINERALS

potassium chlorate at a boiling temperature, the solution beirg
sufficiently dilute to dissolve the salt of lead as fast as it is formed.
In this way a protective coating is avoided and the operation accele-
rated. To the clear and sufficiently dilute solution of the metals,
after filtration from silica, silver chloride, barium sulphate and in-
soluble matters generally, 5 grams of tartaric acid and 5 grams of
citric acid are added and then sufficient caustic soda to approximately
neutralise the liquid, which at this stage should be diluted to not less
than 500 c.c. An ounce of solid caustic soda is next dissolved in the
neutralised solution, which is heated to boiling, and sulphuretted
hydrogen conducted through it until precipitation has taken place
and the supernatant liquid has been transformed into a yellow solution
of alkaline sulphydrate. The object of thus adding the soda in two
operations is to ensure the presence of a known excess of alkali
during the precipitation by sulphuretted hydrogen. The boiling
should be continued (in a porcelain basin) for at least half-an-hour,
and until the sulphides assume a heavy, granular condition and sink
rapidly to the bottom, when the ebullition is discontinued.

If water has been added from time to time to replace that lost by
evaporation, rapid collection and washing of the precipitate are prac-
ticable without rupture of good Rhenish filter-paper.

The precipitate contains the copper, lead, zinc, and iron sulphides,
and also the whole of the gold and platinum contained in the sample;
nickel, if present, demands the previous separation by sulphuretted
hydrogen of the metals forming sulphides insoluble in dilute acids. The
solution contains the arsenic, antimony, tin, selenium, tellurium, &c.

(a.) Treatment of the Precipitate—After two or three washings,
the sulphides of lead, copper, &c., are removed from the paper into a
wide beaker by a jet of water, the filter is burnt, the ash added to
the beaker, and the mixture is covered up and allowed to digest after
addition of nitric acid. Copper and lead salts are formed on warming,
and when the sulphur has become yellow, a sufficient excess of dilute
sulphuric acid is added. The beaker being now uncovered, evapora-
tion takes place, the nitric acid is expelled, and the sulphuric acid
concentrated. Enough water is now cautiously employed to produce
a dilute acid of about 1 part of acid to 8 parts of water, and in order
to ensure the complete separation of lead, the acid is once more con-
centrated and diluted to the same extent. The lead sulphate, mixed
with all the gold and platinum, is next filtered off. These precious
metals may be separated from the lead by solution of the latter in
acetate of ammonium, or by treatment of the mixed precipitate with
a very little aqua regia, which, however, causes a few milligrams of
lead to dissolve. The gold is separated by ferrous sulphate or oxalic
acid, and the platinum then determined by well-known methods. I

CONTAINING THE HEAVY METALS, ETC. 737

prefer treatment with ammonium acetate as the lead sulphate is
thereby purified.

(b.) Treatment of the Filtrate-—This is acidified with dilute sulphuric
acid, and the sulphides, after digestion at a gentle heat to promote
clotting, are separated by filtration and washed. They may be treated
by any suitable method ; the following is one I have applied :—

The sulphides are oxidised by concentrated nitric acid and bromine,
together with the filter-paper if the precipitate cannot be removed
without loss. Subsequent addition of tartaric acid is essential in
presence of antimony. The solution is treated with ammonia, filtered,
and the arsenic precipitated with magnesia mixture, the precipitate
being dissolved in hydrochloric acid and reprecipitated after filtration ;
this second filtrate is mixed with the first as it may contain a milli-
gram or two of selenium or tellurium, but not more. The mixed
filtrates from the arsenic is concentrated and freed from ammonia by
evaporation, and an ounce of oxalic acid is dissolved init. It is then
saturated with sulphuretted hydrogen whilst boiling, which precipi-
tates antimony, selenium, and tellurium, leaving tin in solution. The
tin is recovered by sulphuretted hydrogen, after evaporation with
strong sulphuric acid to destroy the oxalic acid, and weighed as SnO,
after ignition.

The selenium and tellurium are separated from antimony by solu-
tion in aqua regia, evaporation with hydrochloric acid in presence of
potassium or sodium chloride, and precipitation with sulphurous acid.
They are afterwards separated from each other by boiling with
potassium cyanide, precipitation of the selenium by hydrochloric acid,
and reprecipitation of the tellurium by sulphurous anhydride. The
antimony is precipitated by sulphuretted hydrogen, and weighed as
sulphide or as tetroxide.

The following test analyses were made by this method :—

Taken. Found. Taken. Found.
Sn.... 05000 gram 0°5001 gram | 0°2500 gram 0°2507 gram.
Pb.... 0°5000 ,, 04998 _ ,, 05000 ,, 05002 ,,
Au... 01000 ,, 00998 ,, 00100 ,, 00097 ,,

Taken. Found.

05000 gram 04998 gram.

05000 _ ,, 04975,

0°2000 _,, 01985 ,, precipitated as sulphide
0°2000 ” 0°1996 ” [and ignited.

738 BAYLEY: ANALYSIS OF ALLOYS AND MINERALS.

By Mr. J. C. Welch—
Taken. Found,
Pt.... 0°1145 gram \ .
Au... 01000 0°2145 gram (not separated.)
Sn.... 0°2500 02492 __,,
As,0;. 0°2500 0°2484
Pb... 0°5135 0°4887 (accidental loss).
Cu... 0°5200 0°5199
Se.... 0°0800 0°0750

Te ... 0°2000 0°2032
Sb.... 0°2500 0°2527 weighed as Sb,§;.

In analyses like the latter, the precipitates insoluble in sodium
sulphide should be redissolved and reprecipitated to ensure complete
separation of selenium and tellurium.

German Silver.—Although usually almost perfectly free from more
than traces of Sn, Sb, or As, this alloy may be taken as the type of a
substance containing nickel. The solution of 1 gram or so of the
German silver in the smallest possible quantity of nitric acid is mixed
with a few drops of ammonia and one or two drops of sulphuric acid
(the ammonium sulphate promotes the passage of the current), and
the copper then precipitated by electrolysis with two large Daniell
cells, any undeposited traces being precipitated by sulphuretted
hydrogen, and estimated colorimetrically. Ifsulphuretted hydrogen
has been used it is removed by a little bromine-water, and ammonia
is added in slight excess, and then 10 or 12 oz. of redistilled acetic
acid (B.P. 33 per cent.). The volume before adding the acetic acid
should be about 50 cc. The zinc is completely precipitated by
sulphuretted hydrogen in the cold. A pure white zinc sulphide is
thus obtained, and since no fixed matter has been introduced, evapo-
ration to dryness with sulphuric acid, ignition and weighing of the
nickel as sulphate, gives uniformly satisfactory results. Any iron
present must be separated either before or after the evaporation.

Certain commercial alloys are said to contain aluminium, chemical
analysis is usually not capable of detecting this metal, although it is
said to confer extraordinary properties on the alloy. I cannot speak
as to recent samples of American origin.

I have pleasure in acknowledging the careful and skilful assistance
I have received from my assistant, Mr. J. C. Welch, in many analyses

by these methods.

LXIX.—Further Contributions to the Knowledge of Cyanuric Chloride
and other Cyanuric Derivatives.

By Harorp H. Fries.

In a recent communication (Trans., 1886, 314) on the action of
cyanuric chloride on a2-naphthylamine, I described a series of mono-,
di-, and tri-naphthylamidocyanuric derivatives. I have now suc-
ceeded in obtaining a similar series from B-naphthylamine, toluylene-
diamine, and phenylhydrazine.

Before describing these compounds, I would call attention to a
modification of the method of preparing cyanuric chloride which I
have adopted. Gautier (Annalen, 141, 122) suggested that cyanuric
chloride should be made by passing dry chlorine into a solution of
hydrogen cyanide in ether. The vessel containing the ether is
surrounded by ice, dry hydrogen cyanide is passed in until the liquid is
saturated, and then chlorine. Throughout the greater part of this
operation the hydrogen cyanide is in excess, and a large quantity of
a double compound of cyanogen chloride and hydrogen cyanide is
formed. Claésson (Bihang till K. Svenska Vet. Akad. Handlinger, 10,
No. 5) uses chloroform in place of ether. Chlorine is more soluble in
chloroform than in ether, and an excess of this gas, avoided by
Gautier, is found by Claésson to increase the yield of cyanuric chloride
considerably. Now a still better result is obtained if the chlorine be
in excess from the commencement. When this is done the cyanuric
chloride produced is almost entirely free from bye-products.

I lead chlorine into the chloroform first, then hydrogen cyanide
and chlorine together, always taking care to have the latter in
excess. The process is continued as long as there remains any
unchanged hydrogen cyanide, which point is marked by the stoppage
of the evolution of hydrogen chloride. In many cases the cyanuric
chloride separates from the chloroform during the experiment in

beautiful crystals.

Action of Cyanuric Chloride on B-Naphthylamine.

From the analogy between the behaviour of A- and a-naphthylamine
it was to be expected that derivatives of cyanuric chloride would be
obtained with the former compound analogous to those already
produced with the latter. Thus the following compounds were
possible :—

C;N;ClyNH-C,oH; C;N;Cl(NH-C,.H;)2 C;N;(NH-C,.H;)s

Primary 8-naphthyl- Secondary 8-naphthyl- B-Naphthyl-
amidocyanuric chloride. amidocyanuric chloride. melamine.

740 FRIES ON CYANURIC CHLORIDE

Primary B-Naphthylamidocyanuric Chloride.—Solutions of cyanuric
chloride and 8-naphthylamine in dry ether are quickly mixed together
in the proportion of 1 molecule of the chloride to 2 molecules of
f-naphthylamine. A precipitate of 8-naphthylamine hydrochloride
is formed which is removed by filtration and the filtrate evaporated
to dryness. The residue crystallised several times from hot alcohol
gives beautiful, colourless crystals, melting at 154°, of the primary
8-naphthylamidocyanuric chloride, C;N;Cl,-NH°C,.H;.

Experiment.
A.

Theory. c
~ I II.

53°61 5351 —
2:75 2-99

19°24 -

24°40

100-00

Secondary B-Naphthylamidocyanuric Chloride.—In the preparation
of this compound ethereal solutions are mixed slowly drop by drop,
and in the proportion of 4 molecules of 8-naphthylamine to 1 molecule
of cyanuric chloride. The precipitated 8-naphthylamine hydrochloride

is removed as before, and the residue, after evaporation, recrystallised
several times from alcohol. In this way colourless needles, melting
at 178°, are formed of secondary f-naphthylamidocyanuric chloride,
C,N,Cl(NH-C,oH;)..

Experiment.

Theory. r Acc »
——_, I. If. Il.

276 69°43 6911 — —
16 4°02 441 — —
70 17°60 — 1748 —
355 895 — — 918

397°5 100-00

B-Naphthylmelamine.—One molecular proportion of cyanuric chloride
is thoroughly mixed with six molecular proportions of 8-naphthyl-
amine and heated in a closed tube at 150° for about three hours. The
contents of the tube is then washed with hot water and dried. The
crude melamine thus obtained is crystallised from nitrobenzene and
washed with alcohol. Addition of alcohol to the nitrobenzene
promotes the crystallisation of the melamine, which consists of beauti-

ful needles melting at 209°. Its formula is C;N,(NH-C,.H;);.

AND OTHER CYANURIC DERIVATIVES.

Experiment.

Theory.

Cm
(ee, | % II.
78°57 78:36 —
4°76 527. —

16°67 — 1653 16°73

100:00

Action of Cyanuric Chloride on Toluylenediamine.

It seemed interesting to see if a similar series of compounds could
be obtained with toluylenediamine.

Primary Toluylenediamidocyanuric Chloride.—Solutions of cyanuric
chloride in absolute ether and of toluylenediamine in absolute alcohol
are mixed together drop by drop in the proportion of L molecule of
the chloride to 2 molecules of ‘the diamine. The precipitate which
forms is removed, and the filtrate, on evaporation, gives the primary
toluylenediamidocyanuric chloride, C,;N;Cl,,NH-°C;H,NH,, which de-
composes when heated without showing a defined melting point.

Experiment.

Theory.
a
120 44°45

9 3°33
70 25°93
71 26°29

270 8100-00

Secondary Toluylenediamidocyanurie Chloride. — Theoretical pro-
portions of the solutions of cyanuric chloride and toluylenediamine
are mixed and the precipitated hydrochloride removed as in the case
of the preceding derivative. The crude compound is separated from
the filtrate by evaporation, and is obtained in beautiful needles by
recrystallisation from alcohol. These decompose when heated at
about 172°. This derivative has the formula C;N;Cl(NH-C,H,NH,).

Experiment.

Theory. r

re L II.
204 57°38 6721 —

18 5°07 o38
98 27°56 — 27°83
35°55 9°99 _ — 1023

355°5 100°00
VOL. XLIX. 3D

742 FRIES ON CYANURIC CHLORIDE.

Toluidylmelamine.—A mixture of toluylenediamine and cyanuric
chloride is enclosed in a tube and heated at 100° for several hours.
The product is washed with water, then treated with boiling alcohol
and filtered, when the filtrate deposits toluidylmelamine in crystals
on cooling. These are purified by several recrystallisations. They
show no distinct melting point. Their formula is C;N,NH-C,H,'NH;);.

Experiment.

Theory.
a

—
288 65°30

27 6°13
126 28°57

441 10000

Action of Cyanuric Chloride on Phenylhydrazine.

I have also succeeded in producing a series of compounds analogous
to those already described, in which a phenylhydrazine residue replaces
the chlorine of cyanuric chloride.

Primary Phenylhydrazine Cyanuric Chloride-—Ethereal solutions of
phenylhydrazine and cyanuric chloride are mixed together drop by
drop, and the precipitate of phenylhydrazine hydrochloride removed
by filtration. On evaporation, the filtrate yields a crystalline precipi-
tate which is purified by washing with water, alcohol and ether. It
gives no defined melting point, and has the formula

C,N,;Cl..NH:-NH-C,H;.

Experiment.
Theory. r A
a I II. III.

_ ‘
Cy cee 42°19 4192 — —

H,... 2°74 301 — =~

N;... 70 27°34 — 27:28 27:03

Cl... 71 27°73 _ — — 28:02

—— —_ ——_——_-

256 100°00

Secondary Phenylhydrazine Cyanuric Chloride.—Theoretical propor-
tions of the two substances in ethereal solution are rapidly mixed
together, and, after removal of the hydrochloride formed, the filtrate
evaporated to dryness. The residue recrystallised from alcohol gives
needles of secondary phenylhydrazine cyanuric chloride.

C,N,Cl(NH-NH-C,H;)».

SENIER: ACTION OF HEXABROMACETONE ON UREA. 743

Theory.
a,
C,.... 180 54°96
Hyco 86 4°27
N,.... 98 29°97
Cl.... 355 10°84

327°5 100-00

Anilylmelamine.—When the primary derivative is heated in a
closed tube with phenylhydrazine at 100° anilylmelamine is formed,
and is purified by successive crystallisations from alcohol. It has the
formula C,N;(NH:NH-C,H;);.

Experiment.

Theory. ——
a I. iI.

63°15 63°08 _
5°26 5°64 _—
31°59 — 31°91

(i-

100°00

Chemical Laboratory,
University of Berlin,
July, 1886.

LXX.—Further Observations on the Action of Hexabromacetone
on Urea.

By Atrrep SENIER.

In arecent communication (ante, p. 693) I described a series of experi-
ments showing that of the two isomeric cyanuric acids which Herzig
(Ber., 12, 170) claims to have obtained by the action of hexabrom-
acetone on urea, the one —the a-acid—is identical with ordinary cyanuric
acid, whilst in the case of the other—the 8-acid—farther experiments
were needed before a decisive opinion could be formed. These have
now been made. I have prepared several specimens of the B-acid, care-
fully attending to the directions given by Herzig; they were all similar
in appearance, and consisted of crystals unlike cyanuric acid, and
exhibited generally the properties which Herzig observed. On repeated
recrystallisations from water, however, the crystals were found

3 D2

744 SENIER ON THE ACTION OF

gradually to assume the character of the ordinary acid and the other
distinctions one by one to disappear. The properties given by Herzig
as distinguishing the 8 from the a and from the ordinary acid are :—

I. When heated it does not yield cyanic acid.
II. When treated with phosphorus pentachloride cyanuric chloride
is not formed.
III. It does not give Hofmann’s test for cyanuric acid.
IV. The crystals do not contain water.
V. It is more soluble in alcohol than either the («) or the ordinary
acid.
VI. A barium salt of constant composition could not be obtained.

I. Action of Heat.

None of the specimens which I have examined appeared to yield
cyanic acid when heated in small tubes until they had been three or four
times recrystallised from water; when this was done, however, the
odour of cyanic acid was easily obtained, and could be recognised
beyond any doubt.

II. Action of Phosphorus Pentachloride.

By carefully treating the product obtained by heating the @-acid
with phosphorus pentachloride in a closed tube, in accordance with the
directions of Beilstein (Annalen, 116, 357), I have isolated cyanuric
chloride as announced in my previous paper. The contents of the
tube are in the first place distilled at 150°, which removes the phos-
phorus oxychloride formed, ‘the cyanuric chloride remaining with
other products in the retort. The mixture in the retort is extracted
with boiling dry ether, which, on evaporation, leaves. crude cyanuric
chloride. This must now be washed quickly with water, dried in
a@ vacuum, and again extracted with ether. The ethereal solution
thus obtained, when evaporated, gives the cyanuric chloride in crystals,
which are readily identified by their melting point and characteristic
odour.

III. Hofmann’s Test for Cyanurie Acid.

This test is based on the characteristic property of cyanuric acid of
giving a trisodium salt less soluble in warm than in cold concentrated
solution of sodium hydroxide; I have applied it successfully in the
case of all the specimens of the f-acid which I have prepared, even
after the first recrystallisation.

HEXABROMACETONE ON UREA,

IV. Water of Crystallisation.

Contrary to Herzig’s experience, my specimens of the #-acid
when purified by recrystallisation from water and dried by pressing
between paper and subsequent exposure to the air for an hour, contain
water which is given off below 120°, the proportion increasing with
purification. The product of a fourth crystallisation gave the follow-
ing percentage :—

Theory.
(C3N;) (OH);2H,0. Experiment.

21°8 15°8
The same specimen once more recrystallised retained a larger propor-
tion of water. An experiment was made at the same time under
precisely the same conditions. with the ordinary acid, thus :—
Theory. Experiment, Experiment,
(C;N;)(OH),2H,0. ordinary acid. B-acid.
21°8 21:9 19°7

It is clear then that the B-acid when purified does contain water of
crystallisation, and that this is increased by purification, until, as seen
in my last experiment, it amounts approximately to the two molecules
contained in the crystals of the ordinary acid.

V. Solubility in Alcohol.

Herzig gives no determination of the solubility of the B-acid in
alcohol. It is indeed, as.he points out, more soluble in alcohol than
the ordinary or the a-acid, but by successive recrystallisations its
solubility is gradually reduced; thus :—

Ist 5th 6th
crystallisation.. crystallisation. crystallisation.

0°425 0°240 0°156

In 100 grams alcohol
at 22—24°

This greater solubility, therefore, is due to the influence of some
impurity which recrystallisation removes, and is not characteristic of
the B-acid.
VI. Bariwm Salt.
Herzig did not succeed in preparing a definite barium salt from the
B-acid. By carefully following the directions of Woéhler (Annalen, 62,
252), however, I have obtained, as in the case of the a-acid, the

monobarium compound in beautiful needles. It has the formula
(C,N;)(OH).(Oba) + H,0, which requires—

Theory. Experiment.
Barium 32°71

746 FRISWELL AND GREEN: DIAZOBENZENEANILIDE

It thus appears that by simple recrystallisation from water
“8-cyanuric” acid may be so purified that each of the properties
which were supposed to distinguish it from the ordinary acid in turn
disappears. By this means undoubtedly, small quantities of some
impurity is taken away, which, although not sufficient to materially
affect the analytical numbers as given by Herzig is, nevertheless,
enough to modify in a remarkable manner the properties of the acid.
8-cyanuric acid therefore, like the «-compound, is identical with
ordinary cyanuric acid.

Chemical Laboratory,

University of Berlin,
July, 1886.

LXXI.—The Constitution of Diazobenzeneanilide and its Relation to
Amidoazobenzene. II.

By R. J. Friswett and A. G. Green.

In a previous paper (Trans., 1885, 917) we have attempted to show
that the isomeric change of diazobenzeneanilide (diazoamidobenzene)
into amidoazobenzene, supposed to be brought about through the
intervention of aniline hydrochloride, really consists in a splitting up
of the molecule by the acid present into diazobenzene and aniline, and
that these bodies again combine under the altered conditions with
formation of amidoazobenzene. It appeared, however, necessary that
the anilide should always be first formed, for under no conditions
were we able to obtain a direct combination of diazobenzene and
aniline to amidoazobenzene. This is most strikingly shown by the
fact that the anilide can be produced at a temperature above 30° C. in
the presence of a large excess of aniline hydrochloride or of aniline
hydrochloride and aniline, although after standing for a short time it
begins to change into the isomeride. Whilst on the other hand, pure
diazobenzeneanilide free from aniline, if suspended in dilute
hydrochloric acid (1 molecular proportion) is converted in the course
of 24 hours at the ordinary temperature into amidoazobenzene hydro-
chloride, the liquor containing during the whole time some
uncombined diazobenzenechloride.

The results obtained on continuing our experiments quite bear out
our former conclusions, and we are now extending our investigation to

AND ITS RELATION TO AMIDOAZOBENZENE. 747

other diazoamides. The present paper chiefly deals with the constitu-
tion of diazobenzeneanilide, concerning which we believe we are able
to advance some important evidence.

At the end of our former communication we suggested incidentally
that the apparent analogy of diazobenzeneanilide to azoxybenzene was
in favour of the symmetrical formula C;H;N—N-C,H;. In doing so

NH
we unintentionally somewhat misstated the views of Professor V.
Meyer on this point, and now take the opportunity of correcting the
error.

V. Meyer (Ber., 14, 2447, note to a paper by A. Sarauw), whilst
admitting the symmetrical formula to be consistent with all the facts
then known, yet does not regard it as altogether probable, and
considers that the unsymmetrical formula C,H;-N.NH-C,H; will
explain the facts equally well if a migratory power of the H-atom be
assumed (brought about as he considers by addition and subtraction
of HCl).

Contrary to our preconceived idea, our experiments have shown that
the symmetrical formula is untenable, and that the analogy with
azoxybenzene is only superficial. Although azobenzene is converted
by sulphuric acid into oxyazobenzene, this isomeric change appears to
be quite unlike the change of diazobenzeneanilide, for under no
conditions of temperature or dilution could a trace of diazobenzene
and phenol be detected. It alsoseemed possible that if diazobenzene-
anilide had the symmetrical structure it might be produced from

azobenzenedibromide, C,H;-N-N-C,H;, by the action of ammonia: the
Br Br

dibromide seems to exist, although it is very unstable, but on passing
dry ammonia into its solution in chloroform, azobenzene was regene-
rated.

A positive proof in favour of an unsymmetrical structure,
C,H; N..NH-C.H;, was obtained by displacing the hydrogen of the
imidogen-group by alcoholic radicles (the reaction taking place
with extreme ease), and subsequently splitting up the product
C.H,-N.-NRC,H; by an excess of acid.

A monoalkylated aniline was formed together with diazobenzene,
the former of which could not have been obtained from a body of
symmetrical structure. A similar proof was attempted some time ago
by A. Sarauw (Ber., 14, 2443) by displacing the H-atom by CO by
means of carbonyl chloride, but although his conclusions quite agree
with ours, the results he obtained were very unsatisfactory and incon-
clusive, as from want of knowledge of the properties of diazobenzene-
anilide, no precaution was taken to remove the hydrogen chloride

748 FRISWELL AND GREEN: DIAZOBENZENEANILIDE

formed during the reaction; this must have acted on the anilide to
produce diazobenzene, aniline, amidoazobenzene, tarry products, &c.

The symmetrical formula having been excluded, the only two
remaining formule are :—

(a.) C.HyN-NH-C.H,. (b.) C.HyN—N-NH-C,H;.
II

It might fairly be expected that a substance: of the constitution (a)
would when reduced allow the N-atom outside the chain to be removed
in the form of ammonia, and that hydrazobenzene would be produced.
Only alkaline reducing agents.can be used, since acid reducing agents
cause a preliminary decomposition into aniline and diazobenzene, the
latter being then reduced to phenylhydrazine. We have tried the effect
of various alkaline reducing agents, but the anilide is scarcely attacked
by any. In one experiment we dissolved it in absolute alcohol and
added a large excess of metallic sodium, a trace of ammonia was
evolved, but the greater part of the anilide apparently remained
unaltered, and hydrazobenzene was not formed.

Hence there seems little doubt that diazobenzeneanilide has the
constitution C,H,"N : N-NH-C,H; usually attributed to it.

Diazobenzenemethylanilide (Methyldiazoamidobenzene),
C,H, N ; N-N(CH;) C.H,.

A solution of sodium ethoxide'was prepared by dissolving 30 grams
of clean sodium in 300 c.c. of absolute alcohol, and to this a hot solu-
tion of 200 grams of diazobenzeneanilide in 500 c.c. of absolute
alcohol was added. The mixture was of a deep orange-brown colonr ;
nothing separated from it when it was allowed to cool over lime in an
exsiccator, but on exposure to the air it slowly absorbed water and
diazobenzeneanilide separated in beautiful crystals.

When the solution was nearly cold, 170 grams of methyl iodide was
added and the flask connected with an inverted condenser. A
vigorous reaction quickly set in and the mixture boiled violently
without the application of any external heat. After an hour's
cohobation, half the alcohol was distilled off, and water was added to
the residue; the heavy oily layer was separated, washed, and dried
over calcium chloride. The product thus obtained was a dark-yellow,
heavy oil which only became viscid in a freezing mixture and could
not be caused to solidify by any treatment. It decomposed explo-
sively when heated, and did not volatilise with steam, so that it could
not be further purified. The yield of crude diazobenzenemethyl-
anilide obtained was 200 grams.

AND ITS RELATION TO AMIDOAZOBENZENE. 749

Decomposition by Acid.—If the methylanilide is gently warmed with
an excess of hydrochloric acid, diazobenzene can readily be detected in
the solution, but is converted into phenol if the mixture is heated.
100 grams of the methylanilide was gradually added to 200 c.c. of hot
strong hydrochloric acid: when the violent reaction was over, the
solution was diluted with water, made strongly alkaline with caustic
soda, and distilled. The crude methylaniline (yield 40 grams) was
separated from the distillate, cohobated for some time with an alco-
holic solution of carbon bisulphide and a. little potash to remove a
trace of aniline, separated and rectified. It boiled at 192—193°
(uncorr.). Its platinochloride gave on analysis—

Found.

Calculated. for rf ~
(C,H;-NH Me, HCl),PtCl,, 3 II. ITI.
31°11 31°19 31°32

The acetyl-derivative, obtained by heating the base with an equal
weight of acetic anhydride, crystallised from water in beantiful, long,
white needles having the constant melting point 101° (uncorr.).
The melting point of methylacetanilide is given as 102°. The
alkaline residue left after removing the methylaniline by distil-
lation, contains phenol, which after acidifying was distilled off and
isolated by adding salt to the distillate. It was identified by con-
version into tribromophenol, C;H,BrjOH, which crystallised from
dilute alcohol in very fine, white needles, which had the correct
melting point 96° (uncorr.).

Hence diazobenzenemethylanilide is decomposed by acids according
to the equation—

C.Hs'N : N-NMeC,H; + HCl = C,H;-N : N-Cl + C,H;-NHMe.

Diazobenzenemethylanilide (1 mol.) when left in contact with cold
dilute hydrochloric acid (1 mol.), is slowly converted in the course
of 24 hours into a magma of small, dark crystals. This must con-
sist for the most part of the hydrochloride of methylamidoazobenzene,
but we have not yet succeeded in isolating this substance in a pure
state, as it is very difficult to separate from a small quantity of amido-
azobenzene which it appears to contain.

Diazobenzenebenzylanilide (Benzyldiazoamidobenzene),

C.H;'N ‘ N-N(C,H,)C,Hs.

30 grams of sodium were dissolved in 300 c.c. of absolute alcohol,
and a hot solution of 200 grams of diazobenzeneanilide in 500 c.c. of
absolute alcohol was added. The mixture was cohobated for an hour
or two with 140 grams of benzylchloride.

Se err ge

750 FRISWELL AND GREEN: DIAZOBENZENEANILIDE.

On adding water to the alcoholic solution the product was pre-
cipitated as an oil which was washed with hot water; on cooling it
solidified. ,

After one crystallisation from alcohol it weighed 200 grams, which
was diminished to 170 grams by a second crystallisation ; it then melted
at 80° (uncorr.). By further recrystallisation from a mixture of
acetone and alcohol the melting point was raised to 81°, at which it
remained constant. The pure benzylanilide forms beautiful yellow
needles, very soluble in acetone, moderately soluble in alcohol, insoluble
in water. A determination of the nitrogen by Dumas’ method
gave—

Calculated for
C,9H,,N3. Found.

14°87

eee ee ee eeeeee

The substance decomposes explosively on heating. When left in
contact with 1 molecular proportion of dilute hydrochloric acid,
slightly warm, it evidently undergoes change into the isomeric
amidoazo-compound, but the benzylamidoazobenzene could not be
isolated, as the product always separated as a tar.

Decompositien by Acids.—The diazobenzenebenzylanilide is resolved
by an excess of acid into diazobenzene and benzylaniline. If the
substanve is gently warmed with strong hydrochloric acid diazo-
benzene can be readily detected in the solution, and phenol is
obtained if the mixture is boiled.

50 grams of the benzylanilide was added by degrees to 200 c.c. of
nearly boiling strong hydrochloric acid. Part of the benzylaniline
hydrochloride separated at once and the rest crystallised out on
cooling. The acid mother-liquor which contained the phenol was
diluted with water and distilled, and the phenol was separated from
the distillate by the addition of salt. It was identified as before by
conversion into tribromophenol which melted correctly at 96°
(uncorr.). The crude benzylaniline hydrochloride was treated with
alkali and the product distilled. The base formed a pale yellowish,
crystalline solid, which melted at 33° and boiled above 360° under
ordinary pressure. The yield was about 20 grams.

The hydrochloride forms white plates melting at 203° (uncorr.) ; it
is decomposed by water. On combustion with soda-lime it gave—

Calculated for
(C,H;)(C;H;)NH,HCL. Found.

oe cescsccesecece 6:0

The platinochloride formed fine yellow needles melting at 168°
(uncorr.), tolerably soluble in water. On analysis it gave—

THE ESTIMATION OF FREE OXYGEN IN WATER. 751

Calculated for
{ (C,H) (C;H;) NH,HC1},PtCl,. Found.
25°14.

Hence diazobenzenebenzylanilide is decomposed by acids according
to the equation—

C.HsN : N-N(C,H,)C.H; + HCl = C,HsN:N-Cl + (C.H;)(C;H,)NH.

We propose to prepare various other substituted derivatives of
diazobenzeneanilide and of other diazoamides, and hope by applying
this reaction to diazobenzeneparatoluide and paradiazotolueneanilide,
C,H;-N;H°C,H,, to prove whether these bodies are identical as at
present assumed, and to determine their constitution. We hope also

C,H, N;H-C,H;
to investigate in the same manner the twocompounds | ,
C,H, N;H-C,H,
tetrazodiphenyldianilide and bidiazobenzenebenzidide, which according
to our preliminary experiment do not appear to be identical.

In conclusion, we must express our thanks to the firm of Brooke,
Simpson, and Spiller, Limited, in whose laboratories at the Atlas
Works the above investigation was carried out.

LXXIIT.—COMMUNICATIONS FROM THE LABORATORY OF
UNIVERSITY COLLEGE, BRISTOL.

I. The Estimation of Free Oxygen in Water.

By Miss Karuarive J. Wituiams and Professor W. Ramsay.

THE value of Schiitzenberger’s method of determining the amount of
free oxygen in water has been somewhat discredited by the verdict of
those who have attempted to apply it; and in Sutton’s Handbook of
Volumetric Analysis, p. 229, the subject is briefly dismissed with the
remark: “The method of Mohr of finding the oxidising power of
weak alkaline solutions of oxygen upon ferrous compounds, and
subsequent titration by permanganate, as also that of Schiitzenberger
with sodic hyposulphite and ammoniacal copper solution, regulated
by indigo, are by no means consonant with each other, nor with the
accurate gasometric method of Bunsen.”

It was with the object of arriving at some definite conclusion
regarding the accuracy of this method that the experiments to be here
described were undertaken ; and it may be stated in advance that the

752 WILLIAMS AND RAMSAY: THE ESTIMATION OF

method, in our hands, has proved a quick, fairly reliable method, not
difficult in execution.

Schiitzenberger’s description of his process is to be found in the
Comptes rendus, 75, 879, and also in his work on fermentation
(International Scientific Series, 20, 108). We think it may prove of
interest, before proceeding to describe our own experiments,
briefly to recapitulate here the essential parts of the process as
carried out by the author himself; and also the variations in the
process suggested by others who have made use of it.

The fundamental reaction consists in the absorption uf oxygen by
sodium hyposulphite, NaHSO,. This salt, however, need not be
pure. It is easily prepared in an available condition by adding to a
solution of hydrogen sodium sulphite, NaHS9,, an excess of zinc-dust,
allowing the mixture to stand in acorked flask for about five minutes,
and then, when cool, adding excess of milk of lime, and filtering.
The filtered solution may be kept under a layer of benzene in a
stoppered bottle, and diluted as required. To standardise the sodium
hyposalphite, an ammoniacal solution of copper sulphate, containing
4°46 grams of hydrated salt per litre, is employed. On adding the
hyposulphite to the copper solution, the blue colour changes to yellow,
the copper being reduced to the cuprous condition: 10 c.c. of the
copper solution is equivalent to 1 c.c. of oxygen gas, measured at
normal temperature and pressure. The titration is performed in an
atmosphere of hydrogen.

A solution of indigocarmine is also. employed, and its strength
determined by titration with hyposulphite until the blue colour is
destroyed.

Two methods of estimating oxygen were employed by Schiitzen-
berger. In the first, the water containing dissolved oxygen is
placed in a Wolff’s bottle, into which the burettes containing
respectively indigo solution and sodium hyposulphite deliver; after
the air has been displaced by hydrogen, indigo is added in small
quantity so as to colour the liquid, and hyposulphite is then added
until the colour is destroyed. In the second method, warm water is
first placed in the Wolff’s bottle, and about 40 c.c. of indigo is added.
The liquid is then carefully decolorised by addition of hyposulphite.
A known volume of the water to be tested is then introduced, the
oxygen in which restores the colour of the indigo, partially oxidising
it. The colour is again destroyed by addition of hyposulphite, the
volume taken being noted, and from the volume of hyposulphite, of
which the strength has been previously determined by means of the
ammoniacal cupric solution, the percentage of free oxygen in the
water is calculated. The first of these methods, Schiitzenberger
states, estimates only half the oxygen in the water, and he suggests

FREE OXYGEN IN WATER. 753

that the remaining half combines with the water to form hydrogen
peroxide. The second method is therefore in his opinion the only
available one. On this point we shall have something to say here-
after.

A comparison of Schiitzenberger’s method with Mohr’s was next made
by J. Kénig and L. Mutschler(Ber., 10, 2017). The only difference
between their method of experiment and Schiitzenberger’s is a trivial
oue ; they employed a more dilute solution of indigo. They give nine
series of determinations, referring to water saturated with air, to
ordinary well-water, and to water saturated with pure oxygen gas.
Their first seven determinations show fair concordance between
Mohr’s method, which consists in the oxidation of ferrous to ferric
salt, and Schiitzenberger’s; but the two last, referring to water
saturated with oxygen gas, are at variance with each other. In each
case the hyposulphite method gave higher results; and they concluded
therefore that Mohr’s method is more trustworthy. They also agree
with Schiitzenberger that hydrogen peroxide is produced during the
addition of hyposulphite.

These conclusions were again confirmed by K®énig alone (Ber.,
1880, 154).

F. Tiemann and C. Preusse published the results of their observa-
tions in the Berichte, 12, 1768. Not only do they add the water to
be tested to hot water, but they keep the liquid warm (about 40—
45°) during the estimation. In every case three determinations were
made before emptying their vessel, for they had no arrangement to
syphon off the liquid without interfering with the current of
hydrogen. They recommend that the solutions should be of such a
strength that not more than 6 c.c. of hyposulphite is used to
decolorise 10 c.c. of copper solution; and having once for all
determined the ratio between copper and indigo, they do not sub-
sequently standardise their hyposulphite against copper. They
control their results by comparison with Mohr’s method and also by
expelling the gases by boiling the water to be tested, and estimating
the oxygen by absorption. They conclude that Schiitzenberger’s
method gives results concordant with those obtained by gas analysis,
whilst Mohr’s method, in every case but one, gave results lower than
either. They recommended Schiitzenberger’s method for rapid work,
but claim greater accuracy for the direct measurement of gaseous
oxygen.

August Bernthsen (Ber., 13, 2277) again investigated this
process. He employed pure sodium hyposulphite, considering it to
give better results than the impure solution containing sodium
sulphite. He quotes eight series of experiments, in which his copper
solution was standardised against hyposulphite; and he employed

754 WILLIAMS AND RAMSAY: THE ESTIMATION OF

much stronger solutions of hyposulphite than Tiemann and Preusse
recommended ; indeed in one instance, only 2°65 c.c. of hyposulphite
was required for 25 c.c. of copper solution. Some difficulty was
experienced in detecting the true end-point, and the general verdict is
unfavourable to the process. No examples as regards the application
of the method to the estimation of free oxygen in water are given by
him; his object was rather to test a process for the estimation of
indigo.

On reviewing this evidence, it is on the whole not unfavourable to
the process; and the results now to be given justify this conclusion.

The apparatus used by us closely resembled that recommended by
Schiitzenberger. The reaction took place in a wide-mouthed glass
bottle, closed by an india-rubber cork in which five holes had been
bored. Through two of these passed the exit-tubes of burettes
containing respectively indigo solution and hyposulphite, by means of
the usual arrangement these burettes could be filled from stock-bottles
from below by means of T-tubes. The third hole served to intro-
duce the hydrogen froma continuous supply apparatus ; the hydrogen
having been purified by passing over stick potash and through sul-
phuric acid. Through the fourth hole, a doubly-bent tube was inserted,
serving, when drawn up above the level of the liquid in the bottle, as
an exit-tube for the hydrogen, and when pushed down, as a means of
emptying the bottle. A tube provided with a stopcock, and connected
with a reservoir, passed through the fifth hole, and was used in
introducing the water to be tested into the bottle.

The hyposulphite solution was prepared by saturating a strong solu-
tion of 100 grams of sodium sulphite with sulphurous anhydride, adding
20 grams of zinc-dust, and cooling. Milk of lime, prepared from
about 20 grams of quicklime was added, and the resulting solution
was then transferred into the stock-bottle, which was fitted up like a
wash-bottle, the liquid, however, being covered with a layer of
benzene to exclude air. Such a solution is serviceable for at least two
months. This solution was diluted for use until about 7 to 9 c.c.
is equivalent to 10 c.c. of copper solution. The weak solution was
kept in a bottle guarded from oxygen by a layer of benzene and by
bulb-tubes full of potassium pyrogallate ; from this the burette could
be supplied from below as required. Before use, this solution was
always standardised against copper solution or indigo.

The solution of copper was prepared according to Schiitzenberger’s
directions ; the solution of indigo, however, recommended by him was
found to be too strong to give good results. With a strong solution
of indigo, the colour change is very difficult to detect. The most
satisfactory results were obtained with a solution of indigo of such
strength that 10 c.c. of hyposulphite was required to decolorise

FREE OXYGEN IN WATER. 755

20 c.c. Some experiments were made using warm water; but it was
soon found that identical results were obtained, only after a some-
what longer time, at the ordinary temperature, and our experiments
were therefore made without application of heat.

The experiments were conducted as follows :—Having expelled all
air from the bottle by a current of hydrogen, the water to be tested
was admitted through the bulb-tube ; a few c.c. of indigo was added
to tinge the water, and then hyposulphite was added from the
burette until the blue colour changed to yellow. On standing for
some minutes, it again changed to blue. More hyposulphite was
added from time to time until the solution remained permanently
yellow. The indigo having previously been compared with hyposul-
phite, the amount of the latter equivalent to the small quantity of
indigo added was deducted from the total amount used; and the
remainder was assumed to have reacted with the free oxygen in the
water. This is the first method given by Schiitzenberger.

The other method employed was to add a considerable amount of
indigo to about 100 c.c. of water in the bottle, and to destroy its
colour with hyposulphite. The water to be tested was then added,
and the resulting blue colour again destroyed. In this case the
amount of oxygen is calculable directly from the amount of hypo-
sulphite taken.

Results.

We first give instances of the application of Methods I and II,
with full details, so that some idea may be gained as to the concor-
dance of the individual results with each other.

1. Standardising solutions :—

Sodium hyposulphite and copper solution. Hyposulphite and indigo solution.
Bie sles

Copper solution taken. Hyposulphite used. Indigo taken. Hyposulphite used.
10 c.c. 9°31 c.c. 6 c.c. 1°34 c.c.
10 9°26 S ws 1:50 ,,
10 9°39 S 1°52 ,,
10 9°40 O is 1°40 ,,
10 9°36 Mean 6 ,, 1°44 ,,
Mean 10 ,, 9°34 ,,
And 10 c.c. indigo require 24 c.c. hyposulphite.

2. Determinations of free oxygen in water, according to Method I.
This shows the first stage. It is not always easy to stop accurately
when the first reaction is complete, hence the less concordance of the
results of the different series. The water used was ordinary Bristol
tap-water, shaken with air.

756 WILLIAMS AND RAMSAY: THE ESTIMATION OF

Series I:—
Hyposulphite used
Water taken. Indigo. Total hyposulphite. for oxygen.
300 c.c. 4 c.c. 9°73—1-00 8°73 c.c.
4 ,, 9°96—1-00 8:96 ,,
4°25 c.c. 9°78—1-06 8°72 ,,
4 c.c. 9°87—1-00 8:87 ,,

Mean 8°82 c.c.

The strength of the solution of indigocarmine was such that it
required 1 c.c. of hyposulphite to decolorise 4 c.c.of indigo. 1000c.c.
of water from the-above results would require 29°4 c.c. of the hypo-
sulphite; and as 9°41 c.c. of hyposulphite was equivalent to 10 c.c.
of copper solution, each c.c. of eopper solution being equivalent to 1 c.c.
of oxygen at normal temperature and pressure, the total amount of
oxygen per litre is:3°124 c.c.

Series II, III, and IV, Bristol ‘tap-water, shaken with air. The
equivalent of the added indigo in hyposulphite has been subtracted.

Hyposulphite used for oxygen.

Water taken. ft. IIT. IV.
250 c.c. 6°28 :c.c. 8-08 e:e. 7°83 c.c.
6°20 ,, 8°06 ,, 774 ,,

601 ,, 801 ,, -

— 818 ,, —~

Mean 6°16 c.c. 8°08 c.c. 7°78 c.c.

Series IT. 7°397 c.c. NaHSO, = 10 c.c. copper solution. 1000 c.c.
water required 24°64 -c.c. NaHSO,; therefore free oxygen
= 3331 c.c.

Series III. 8°94 c.c. NaHSO, = 10 c.c. copper solution. 1000 c.c.
water required 32°32 c.c. NaHSO,; therefore free oxygen =
3°619 c.c.

Series IV. 9°34 c.c. NaHSO, = 10 c.c. copper solution. 1000 c.c.
water required 31:12 c.c. NaHSO,; therefore free oxygen =
3°112 c.c.

Second stage of Method I.
Bristol tap-water shaken with air was again employed. Series I,

II, and IIT.

FREE OXYGEN IN WATER.

Hyposulphite used for oxygen.
A.

: ae ~
Water taken. § ah. III.

250 c.c. 12°72 c.c. 10°27 c.e. 9°92 c.c.
12°43 _,, 10°36 _,, 9°70 ,,
12°64 ,, 10°52 ,, 10°04 ,,
12°45 _,, 10°05 ,, —

12°75 _,, os --

9°88 c.c.

12°598

Mean

Series I. 9°34 c.c. NaHSO, = 10 c.c. copper solution. 1000 c.c.
water required 50°232 c.c. NaHSO,; therefore free oxygen =
5°371 c.c.
Series IT. 7°397 c.c. NaHSO, = 10 c.c. copper solution. 1000 c.c.
water required 41°20 c.c. NaHSO,; therefore free oxygen =
5°570 c.c.
Series IIT. 7°21 c.c. NaHSO, = 10 c.c. copper solution. 1000 c.c.
water required 39°52 c.c. NaHSO,; therefore free oxygen =

5°480 e.c.
3. Method II. The following results were obtained with Bristol
tap-water shaken with air.

Series I and II.

Hyposulphite used to decolorise

indigo.
C as —
Water taken. 4 II.
100 c.c. 5°06 3°86
511 3°68
5°20 3°62

j 3:97
—_— 3°82

Mean 5°236

Series I. 9°41 c.c. NaHSO, = 10 c.c. copper solution. 1000 c.c.
water required 52°36 c.c. NaHSO,; therefore free oxygen =
5°564 c.c.

Series II. 7:21 c.c. NaHSO, = 13 c.c. copper solution. 1000 c.c.
water required 37°7 c.c. NaHSO,; therefore free oxygen =

5°229 c.c.
It will be seen that the first method gave, when pushed to its last
stage, as a mean result of three series ,5°474 c.c. of oxygen per litre ;
VOL. XLIX,. 3 E

758 WILLIAMS AND RAMSAY: THE ESTIMATION OF

this number agrees well with that obtained by the second method, as
a mean of two series, 5°396 c.c.

It appeared desirable to check these numbers by a direct measure-
ment by gas analysis of the total oxygen expellable from water by
boiling. The carbonic anhydride was removed with caustic potash,
and the oxygen with cuprous chloride. The following results were
obtained with Bristol water shaken with air :—

1. Free oxygen....
2. Free oxygen.... 5°26 Mean 5°402 c.c
3. Free oxygen.... 5 ii
4, Free oxygen....

Mean of other methods, 5435 ¢.c. It is thus seen that Schiitzen-
berger’s method may be regarded as a fairly accurate process for
determining free oxygen in water, whichever way be adopted for its
determination.

It was suggested by Schiitzenberger, that the discrepant results
obtained by his first and by his second methods depended on the
formation of peroxide of hydrogen; he stated that only half the
oxygen was estimated by the first method, and he accounted for this
by the supposition that the remaining half of the oxygen was con-
verted into hydrogen peroxide. Konig (Ber., 13, 154), on comparing
Mohr’s and Schiitzenberger’s processes, found too high results by the
former; he supposed that the discrepancy was caused by the pre-
sence of hydrogen peroxide, which, he stated, acts on ferrous sulphate,
but not on sodium hyposulphite.

Now it will be seen, on comparing our results by Schiitzenberger’s
first and second methods, that the mean result of four series of
experiments with the first, is 3°2961 c.c. of free oxygen, or rejecting
the third series which gave high results, 3°189 c.c. ; and on taking the
ratios of the mean of Method I to that of Method II, the number
1°66 is obtained ; or if Series III be rejected, 1:7. It is therefore
clear that the ratio of the oxygen estimated by the first half of the
first method to the total oxygen is not, as stated by Schiitzenberger,
1: 2, but 3: 5, approximately. But it is difficult to devise an equa-
tion which will in a rational manner account for this partition of
oxygen, It may possibly be caused by some partition between indigo
and hyposulphite, for which the conditions of temperature and dilu-
tion were favourable to the proportion found.

In spite of Schiitzenberger’s statement, for which there appeared
little foundation, that hydrogen peroxide is not attacked by sodium
hyposulphite, we thought the experiment worth a trial, Kdnig also
confirms Schiitzenberger’s remark, but our results do not bear out
their conclusions.

FREE OXYGEN IN WATER. 759

A dilute solution of hydrogen peroxide was prepared, and stan-
dardised with weak potassium permanganate containing 2°661 grams
per litre. Although hydrogen peroxide is heavier than water, yet
such a dilute solution has so nearly the same specific gravity that the
difference may be neglected. The mean of four very concordant esti-
mations proved that the solution of peroxide contained available
oxygen equal to 221°4 c.c. per litre. It was immediately tested with
hyposulphite by Method I, with the results which follow :—

Solution of peroxide taken. Hyposulphite employed.
100 c.c. 6°75 c.c.
- 6°84 ,,
" 700 ,,

Mean 6°86 c.c.

9°41 c.c. of NaHSO, = 190 c.c. of copper solution. 1000 c.c. of
peroxide solution required 68°6 c.c. of NaHSO,; therefore
oxygen estimated = 7°28 c.c.

On standing for some minutes, however, the solution again regained
its colour, and required further addition of 13°59 c.c. of hyposulphite
before a permanent discharge of colour was produced. This is equiva-
lent to a total of 204°5 c.c. per litre, and oxygen estimated = 20°67 c.c.

The same solution of peroxide was then tested by Method II.

Solution of peroxide taken. Hyposulphite employed.
100 c.c. 18°99 c.c.
m 18°39 ,,
18-58 ,,

”

Mean 18°68 ec.c.

4°87 c.c. NaHSO, = 10 c.c. of copper solution. 1000 c.c. of
peroxide solution required 186°8 c.c. of NaHSO,; therefore
oxygen estimated = 38°36 c.c.

A fresh solution of peroxide was prepared, and as a mean of five
very concordant estimations with potassium permanganate, was found
to contain available oxygen equal to 215°76 c.c. per litre. The same
solution was tested with potassium iodide and sodium thiosulphate,
and was found to contain 214°U6 c.c. of available oxygen per litre. It
was titrated with hyposulphite by both methods, with the following
results :—

760

Solution of

peroxide taken.

Method I.

THE ESTIMATION OF FREE OXYGEN IN WATER.

Intermediate stage.

™
Final stage.

Method II.

100 e.e. 6°71 c.c. 17°93 c.c. 37°01 c.c.
- 6°53 ,, 18°23 _,, 36°73 ,,
: a 18-04 ,, 40°39 ,,
n 616 ,, 17:90 ,, ae

Mean 6°46 c.c. 18°02 e.e. 38°04 e.ce.

8:78 c.c. NaHSO, = 10 c.c. copper solution. 1000 c.c. of peroxide
solution required for intermediate stage, 64°6 c.c. of NuaHSO, =
7°35 ¢c.c. of free oxygen; and in final stage, 180°2 c.c. of
NaHSO, = 20°52 c.c. of free oxygen.

By Method II, 1000 c.c. required 380°4 c.c. = 43°20 c.c. of free

oxygen.

A third solution of peroxide was prepared, and contained available
oxygen equal to 159°06 c.c. per litre. Two estimations by Method I
showed that in the final stage 15°03 c.c. of this oxygen was extracted.
After standing all night, this amount was increased to 16°60 c.c.

These determinations show that peroxide is partially estimated by
hyposulphite. The strength of the first and second samples of
peroxide was nearly the same, and the results by Methods I are
nearly identical. But the third sample of peroxide was more dilute,
and appears therefore to be more stable. It is noticeable that the
results by Method I pushed to its concluding stage, and by Method II,
are here by no means identical. We have here also, doubtless, an
instance of partition of oxygen between indigo-white and hyposul-
phite, constant for the particular dilution and temperature employed.
It may also be concluded that on diluting hydrogen peroxide with
water some decomposition takes place, for after standing over night,
the amount of free (or available?) oxygen had slightly increased.

It seemed to us that it might be interesting to ascertain the loss
of oxygen suffered by water on boiling briskly. This was suggested
by the very great difficulty experienced by one of us in expelling all
the air from water by boiling, even in a vacuum.

After boiling for half-an-hour, Bristol tap-water contained still
2-09 c.c. of free oxygen per litre; and after an hour’s boiling, the
amount was still 0°45 c.c. When frozen by evaporation by a Carré’s
pump, and then thawed, the water still contained 1°95 c.c. of free
These experiments show how difficult it is to remove

oxygen per litre.
oxygen gas from water.

In conclusion, we would state that we are confident that the pro-
cess of estimating oxygen, recommended by Schiitzenberger, is fairly
accurate and rapid; and that both methods described by him give

RICHARDSON : DETERMINATIONS OF VAPOUR-PRESSURES. 761

identical results, when the reactions are allowed to reach their final
stage. The second is the more rapid, and is consequently preferable.

Addendum.—Since writing the above, our attention has been drawn
to a paper by Dr. Dupré, published in the Analyst, 11,156. We agree
with him that the estimation of free oxygen in a sample of water,
saturated with air, forms a reliable method for standardising the hypo-
sulphite solution. We cannot, however, agree with him in his con-
clusion that when Schiitzenberger’s first method is used, the end-point
of the first reaction indicates that exactly half the oxygen in the
water has been estimated. Dr. Dupré gives no analytical details to
confirm this statement; and those adduced by us in the foregoing
paper show conclusively, we think, that the hyposulphite required for
the partial reaction does not bear to that required for the completed
reaction the proportion 1 : 2, but 3:5. Our experiments, however,
bear out Dr. Dupré’s, in the more important conclusion that the final
stage of Method I gives results identical with those of Method II.

LXXIII.—Determinations of Vapour-pressures of Alcohols and Organic
Acids, and the Relations existing between the Vapour-pressures of
the Alcohols and Organic Acids.

By Artuur Ricnarpson, Ph.D., University College, Bristol.

Apparatus Used.—The vapour-pressures of the alcohols and acids
were determined by means of the apparatus described by Drs. Ramsay
and Young (Chem. Soc. Trans., 1885, 42). The following alcohols and
acids were examined :—

Methyl Alcohol, obtained from Kahlbaum as free from acetone,
was distilled after standing 24 hours over calcium oxide, the last
traces of water being then removed by treatment with small quantities
of metallic sodium and subsequent distillation. It boiled at 658°
at 760 mm. .

Ethyl Alcohol.—A nearly pure specimen was provided by Dr.
Ramsay, the last traces of water being removed by distillation after
treatment with metallic sodium. It boiled at 78°2° at 760 mm.

Normal Propyl Alcohol (pure from Kahlbaum) was distilled
after treatment with metallic sodium, as in the previous cases; the
fraction which distilled over at 97°2° at 746°4 mm. was taken as

pure.

VOL. XLIX. oF

RICHARDSON: DETERMINATIONS OF

Experimental Results.

The numbers representing the vapour-pressures of the alcohols and
acids at different temperatures are given in the following tables; in
all cases two series of experiments were made for each substance in
order to check the results obtained.

Vapour-pressures.

— 8°3 17°3 |—10°1| 15°05 4°2 16 89 |- 3°3 | 10°23
—7°3 19°22 |— 7°3 18 *32 4°8 17 “94 +73 22°39
— 3°3 24°33 — 6°3 19-09 8°8 23°31 "13-6 30 67
—2°3 25 °23 — 6°3 19°44 9°8 24°51 19°8 45°22
+07 32°06 — 5°3 21°24 10°7 | 25°80 23 *5 56°25
2°23 35°15 — 4°8 20 *54 17°2 | 88°57 26°8 68 -96
4°2 38°30 — 17 26 *92 21°8 50°72 29°6 80°61
6°25 43°88 + O°2 28 ‘67 22 °6 51°96 30 °7 83 °86
8°5 18 “97 4°7 86°89 23 °1 54°86 2°38 | 95°10
1-2 | 55°82 7°7| 48°87 | 25°3 | 62°48 | 34°8 | 104-81
2°0 59°84 12 °2 58°48 27°3 71°11 87°3 | 119°97
i°7 77°99 14°7 68°72 30°7 | 81°86 42 °2 155 °59
7 108 -96 20°7 94°03 35 °3 107° 22 43°8 | 165°41
3 133°55 23 °7 | 112°19 36°6 | 114°59 46°5 | 189°56
2 | 155-34 25°7 | 124-85 40-1 | 133°41 | 50-05 | 226-99
3 183°77 31°7 | 158 °07 43 °1 162 *96 52°8 257 °52
2 235°13 | 36°2 | 202 °83 47 *2 199°18 55°50 | 258°49
7 | 289°84 | 46°6 | 332-37 49°6 | 222°87 59°80 | 349°02
2 370°26 | 54°7 | 459°22 53°7 269 °83 65°80 | 451°48
7 500° 20 58°7 | 552°47 57°3 320 *87 68°80 | 513°76
9 | 584-24 66°45 754°9 61°0 | 370712 | 71-10 | 569-64

7 3° 642°

Isobutyl Alcohol, supplied pure from Kahlbaum, was fractionated ;
after treatment with sodium the greater portion -passed over at 107°7°
at 755 mm.; this fraction was used.

Isoamyl Alcohol.—A sample, provided by Dr. Ramsay, was fraction-
ated ; after treatment with sodium that portion boiling constantly at
130°1° at 7€0 mm. was taken as pure (p. 764).

?lycerol.—A sample was taken, distilled ina vacuum, and the distil-
late collected when the boiling point was constant; this portion was
considered free from water.

VAPOUR-PRESSURES OF ALCOHOLS AND ORGANIC ACIDS. 763

Formic Acid, obtained from Kahlbaum, was distilled, and the
fraction which boiled constantly at 101:3° at 760 mm. was used.

Acetic Acid.—A pure specimen was provided by Dr. Ramsay. It
boiled at 118°5° at 760 mm. (p. 765).

Propionic Acid.—A specimen obtained from Kahlbaum was fraction-
ated, and the distillate which passed over at 140°3° at 760 mm. was
taken as -pure.

Isobutyric Acid was fractionated, and the portion which boiled con-
stantly at 153° at 760 mm. was used (p. 766).

Isovaleric Acid.—An aqueous solution was distilled, and the fraction
boiling above 150° was repeatedly shaken with phosphoric pentoxide ;
on again distilling, the greater portion passed over at 174°9° at
750 mm.; this fraction was taken as pure (p. 767).

Vapour-pressures—continued.

Normal propy] aleohol. Isobutyl alcohol.

Ist experiment. | 2nd experiment. Ist experiment. 2nd experiment.
|

| }
|
Temp. | Pressure | Temp.
in mm.

inC’. | nmm. | mc. | in mm. in C°. | in mm. in C

11 ‘212
21°
26°
33 °
44°
55°
71°
85°
103°
123°
124°
134°
166°
182°
204 -O§
229°
259°
297 587
347 467
399° 35
428 ‘94
468 °978
527 *363
575 651
630° 890
680 °830
755 °114

25 °
35
37

32
46°
52
57°
61
66
67
69
74
77°
78
80
81
83°
85

1145 | 8
14° 10-96
23° 18°
28 24-96
33 34
35° 39°
36° 40°
33° 45°%
39 47 «
41° 54° 8!
44° 65:
48° 84°33
53°9 | 106-6
55 119°2
56°5 | 129°
59°7 | 143°
62 163 °
66 214-7:
68°7 | 220°
74° 287°
79°2 | 358-
85°7 | 477°
566° 25
745°

DONTE Nd ATS

NOOO TN PP
ee or . .
SISO Om w

BD bo AT DO UAT DD tT tA S SAT Aa

AON Hod

NEON WTATTAT AGA ATO NO
wd ede

DOAWHSH
Or or or
. . , nr t x
DWISHRONDOOTONDHHOSOHDMHOOTO OK.
eee ODDHSODWHODWSGOESS

aS
oS
N=4
@

RICHARDSON : DETERMINATIONS OF

Vapour-pressures—continued.

Tsoamy] alcohol. Glycerol.

lst experiment. 2nd experiment. lst experiment. 2nd experiment.

27 35 °8 118 °45 , 2° ‘597
‘97 46°5 , 118 “75 ; 27 °$ *746
‘87 47°8 ‘ 120°77 ; 5°5 "165
58 58° 122 °92 : 5 *340
*93 65 123 85 ‘ 5° ‘285
2°87 67° 127 *91 : ‘ *659
“60 70 130° . : ‘417
‘98 73 134° ‘ 19°7 059
93 74 137° : “5 683
“92 76° 138° . ' 512
‘08 78 139°
07 81 141°
15 82 141°
"04 84 143°
“72 85 147°
“71 151°
“96 161° 25
‘0 162°
“67 171°
‘80 ' 90° 172°
*20 ' ‘6 183°
21 ; ' 195°
“81 . “BS 200 -
“40 . . 201°

“92

Swmad
ur

oro or

IARSHSSAS HOw
AR

| 258 °627
266 °85
347° 092

WHAReEDSWMADHS

VAPOUR-PRESSURES OF ALCOHOLS AND ORGANIC ACIDS. 765

Vapou r-pressures—continued °

Formic Acid.

Acetic Acid.

lst experiment.

2nd experiment.

1st experiment.

2nd experiment.

Temp.
in C”.

Pressure
in mm.

Temp.
in C’.

Pressure
in mm.

Temp.
in C”.

Pressure

Pressure
in mm.

RTO ATTN ATE

151

| 529

| 562°
“89
76
728°

| 648
677

‘9
185°
245°
318°
391°
467°
03

39
2

72
23
47

18
22
29°
39
+t
57°
64°
68°
71
75°

WIR OD He TT Er

28 °657
34° 588
47 °844
73 262
82°97
89
423
"22
“90
088

3rd experiment.

Temp.

in C’.

Pressure
in mm,

bo bo 03 547-3 0d WD DDE AT

13°46
14°60
13°46
17°44
20°93
32°94
48 °33
65 *02
84°82
99 -165
152 ‘96

DE ADRAM Eom AG

KODAAMN AT

RICHARDSON: DETERMINATIONS OF

Vapour-pressures—continued.

Propionic acid. Isobutyric acid.

lst experiment. 2nd experiment. 1st experiment. 2nd experiment.

Temp. | Pressure} Temp. | Pressure} Temp. | Pressure | Temp. | Pressure
inC. | inmm. | inC. | inmm. |] inCe | inmm. | inC.” | inmm.

‘996 “0% 5 873
‘097 | § “366
“193 7 *2 ‘90
2°798 | 9°6 ‘73
"584 35° 3°25
3°787 92 °é ‘7
2°158 98° 97°79
°436 4°8 27°90
‘78 of 4 ‘00
> 369 23° 279 04
‘09 26° *67
29°75 | 131°: *84
‘09 =| 136°05 | ‘19
*237 | 143° 552 °24
5°03 | 153° 51°27
3°66
“76

15°
16°
20
28°
31°
31
35
39
46
55
69
77
83°
87°
100
108
113°
120°
125° 5°59
2° 03
135 °7 350 * 39° | 522 °03
140°3 30 ° 43° 01°49

ou
oro

OO or cr & dO bo bo

SINT RS STR to GATE ATA S
—
DODO ODODERODOMANASSH

2
7
7

‘0

7
7
2
7

7
7
3
4

\ Bb gh oo ob

~
ow
~

VAPOUR-PRESSURES OF ALCOHOLS AND ORGANIC ACIDS. 767

Vapour-pressures—continued.

Isovaleric acid.

lst experiment,

Temp. Pressure
in C.° in mm.

2nd experiment.

3rd experiment.

Temp.
in C.°

Pressure
in mm.

Temp.

Pressure
in mm.

3
o°4
8
9°9
°2
9
< *4
‘8
‘8
‘8
9
534,
‘0
5°8
8
‘8
6
‘3
‘3

51°

58
62

73°5

89°:
92°

96°
101°:
106 *2
112°

114°
116°7

133°

143°
148 *6
150°
153°
158 *%
162°
165°
171°:
176°0

56 * 2:

s7 <3 to I

65 °4

84°12

93 *3:

93° 8%

112 “6%

124 °05

138 *2:

‘741
3 °335
‘932
‘977
‘725

9 955
*914
291
372
9-867
‘087
5°977
€ *45
5 “891
3 °805
2°273
*482
8-799
2-298
75°159
218-114
>°192
2-085
“386
9-531
‘039
2°774
‘877

73 +252
“06

587

3°388
“714
« “99
‘88
"149
“588
727
2°595
785
‘01
552
*262
443
*310
‘716
"95
“611
"53
‘91
"52
“7

“99

From the construction of the apparatus used for making the
vapour-pressure determinations, it is obvious that errors arising from
the vapour becoming superheated are entirely avoided, for even if
the vapours should become superheated, the liquid in contact with
the thermometer bulb must be at the true boiling point, since it has
a free surface of evaporation; that the temperature was sufficiently
high was insured by keeping the temperature of the bath at least
30° higher than that of the evaporating liquid. The numbers given
also represent the true vapour-pressures of the substances experimented
on, for it has been shown by Regnault and others that the statical and

768 RICHARDSON : DETERMINATIONS OF

dynamical methods of measuring vapour-pressures give identical
results; as, however, this identity has been denied by Kahlbaum
(Siedetemperatur und Druck, Leipzig, 1885), the vapour-pressures of
methyl and ethyl alcohol as obtained by the two methods are compared
in the following table :—

Methyl alcohol. Ethyl alcohol.

|
|

Regnault. Richardson. Regnault. Richardson.

21°
34
41°
47
52
56
59
62
65°
68°
70

G
°
fo)

22
34
42
48
52
56
60°
62°
65
68
70

or.

ddonéadduEe
RAESCEASSES
WBAROWOHHOW OW
SCHOUAKSNWEBOD

The vapour-pressures of some of the acids of the acetic series have
been determined by Landolt (Annalen, Sup. 6, 129), but I believe his
results are untrustworthy for the following reasons :—

(1.) In the case of the higher acids, the observed boiling points do
not agree with those calculated from his formule; for example:
Butyric acid; observed boiling point 163° at 765°3 mm.; calculated
160°8° at 760 mm. Valeric acid; observed boiling point 175°6° to
176°1° at 766°2 mm.; calculated 173°7° at 760 mm.

(2.) In no case does Landolt appear to have boiled the mercury in
his barometer tube; unless this is done it is impossible to insure
absence of air in the tubes filled by such a method as he adopted.
Landolt believed that errors due to this cause were eliminated by
tilling both the barometer and the experimental tubes in precisely the
same way ; it is extremely probable, however, that air adhering to the
glass in the experimental tube would be carried up by the liquid as it
rose through the mercury on being introduced from below, whereas in
the barometer tube the air would remain for the most part adhering
to the sides of the tube.

(3.) If Landolt’s results for valeric acid are closely examined it will
be seen—

(a.) That the boiling points of the two samples which he employed
were almost identical.

(b.) That the vapour-pressures of the two samples at 40° were

* = ee Se © a

VAPOUR-PRESSURES OF ALCOHOLS AND ORGANIC ACIDS. 769

respectively 18°9 mm. and 12 mm., corresponding to a temperature
difference of about 14°.

(c.) That the vapour-pressures at 178°2° (slightly above the boiling
point) differed in the two samples by 49°5 mm., equivalent to a
difference in temperature of about 2°.

Now it has been proved that vapour-pressure determinations by the
dynamical and statical methods are identical when the substance under
examination is pure, and the fact that the boiling points of Landolt’s
two samples were almost identical whilst the vapour-pressures at the
boiling points were widely different, can only be explained in one of
two ways, either the substances were impure, or the determinations
of the vapour-pressures were inaccurate (this remark applies to 1).

(4.) It has been shown with more than 50 substances that, when
compared with water, the value C in the equation R’ = R + C(t’ — t)
is a constant through wide limits of pressure; the only observed excep-
tions are mercury, as determined by Regnault, and acetic, propionic,
butyric, and valeric acids, as determined by Landolt; owing to this
apparent anomaly in the case of mercury and to the unsatisfactory
nature of Regnault’s experiments with this substance, the vapour-
pressures of mercury have been redetermined by Ramsay and Young,
who find that mercury does not really form an exception to the above
law. If the vapour-pressures of the acids given in this paper are
correct, then these substances also conform to the law, whereas if
Landolt’s results are correct they are the only known exceptions to it.

Observed Relations between the Vapour-pressures of the Alcohols
and Acids.

These substances were compared amongst themselves in order to
ascertain whether a relation existed between them similar to that ob-
served by Ramsay and Young (Phil. Mag., Dec., 1885) between chloro-
benzene and bromobenzene and between chloride and bromide of ethyl
and other nearly allied compounds, where it was shown that the
ratios of the absolute temperatures corresponding to equal vapour-
pressures remain constant for very wide ranges of vapour-pressure.
The acids and alcohols were also considered with reference to the
further relation which exists between compounds not necessarily allied
in properties and composition, for Ramsay and Young have shown
that the ratios of the absolute temperatures of any two substances
corresponding to an equal vapour-pressure are either constant for all
vapour-pressures, or else the ratios increase or decrease with rise in
temperature by an amount which is directly proportional to the rise
in temperature.

In the subjoined table, the ratios of the absolute temperature of the

770 RICHARDSON : DETERMINATIONS OF

alcohols and acids are given for vapour-pressures between 50 mm. and
750mm. It will be seen that in the case of methyl and ethyl alcohol,
the ratio of the absolute temperature at 50 mm. is 1: 1°045, diminish-
ing with increased vapour-pressure to 1: 1°035 at 700 mm. ; a decrease
is observed in the value of the ratios between ethyl und propyl alcohol
from 1: 1°062 to 1:1°051. In the case of propyl and isobutyl alcohol,
however, the variation is much smaller, the ratio at 50 mm. being
1: 1:027, at 700 mm. 1: 1°029, whilst the value of the ratio of isobuty]
and isoamyl alcohol is practically a constant for all vapour-pressures
between 50 mm. and 750 mm.

Similarly the ratios of the absolute temperatures of formic acid and
acetic acid vary considerably, being at low vapour-pressures 1 : 1059,
falling as the pressure rises to 1:1°049; so also with acetic and pro-
pionic acids, a fall in ratio from 1: 1°071 to 1: 1°054, accompanies the
rise in vapour-pressure. On ascending the series, propionic and
isobutyric acids are much more nearly constant in the values of their
ratios at different vapour-pressures, being at 50 mm. 1: 1°030, and at
750 mm. 1: 1°037, the value of the ratios approaching a constant for
all vapour-pressures between 50 mm. and 750 mm., in the case of
isobutyric and isovaleric acids.

It would appear, therefore—

(1.) That the ratios of the absolute temperatures of the lower
alcohols and acids diminish with increased vapour-pressure.

(2.) That among the higher alcohols and acids in the series the
value of the ratios tends to become a constant number for vapour-

pressures between 50 mm. and 750 mm.

771

6¢°0-T 620-T 1cO- T 9£0- T
620- T 620-1 $co-T | ego. I
6°20: T 620-T £cO-T ceo. T
6¢0-T 6z0- ¥c0-T 9€0- T
6°0- T 820. ¥¢cO- T 9€0-T
660. T §Z0- 9¢0- T 980. T
6°20. T $Z0- 920. T ceo. T
6¢0-T 8Zz0- sco. T ego. T
L¢0-T 8zu: 8¢O0-T 980.T
6¢0- T Lz0. 620-1 | ggo0.T
660. T 9Z0- 6c0: T Lg0.T
090. T 9Z0- 60. T | L80.T
6¢0.-T | L&O- 090- I . | ZO. T
8c0-T | §20-T 040-1 | Zto-T
620. T 420-T | Zy0- 1 cro. T

ee Oe Oo Oe Oe Oe

“splooy *sfOqooulry

‘Wu OGL Pun OG uaamjaq sainssoud-unodn4 sof spwoy pun spoyoy fo ainzosadway, aynjosqy ay} fo ouny oy} fo uosiuvdmog

S
oO
<4
o
=
Aza
<j
ids)
[==]
co)
Q
Z
<j
nm
=)
2)
=e
°
2)
4
=<
fy
io)
MN
<3]
[a]
=)
mM
7A
<>}
(om)
a
om
=)
Oo
A,
na
-

RICHARDSON : DETERMINATIONS OF

Acids—continued.

Iso

Iso
C;H,-COOH to C,H,-COOH.

Tso
CH;-COOH to C,H;-COOH.|C,H,;-COOH to C;H,-COOH.|

053
054
056
056
053
053
052
“054
051
‘O51
051
051

‘037
035
031
030
“031
033
*033
“033
“032
031
031
031
“032 051
032 ee *052
031 o* 1-053

‘071
‘072
‘069
‘069
067
*062
‘061
‘068
061
‘060
059
059
057
056
*054

rt Fd fet feed fd fet freed fed fed fd fe ed fed fet
Prt tt ped ed fd fe fet fe ed fed fed fe fed
al aol aol ated ol ll

a |
|
|

Additional proof of the truth of the second relation is afforded by
comparing the vapour-pressures of these compounds with those of
water, that substance being taken as the standard, since the determi-
nations of the vapour-pressures of water between 0° and 100° (i.e., up to
one atmosphere) are more trustworthy than those of any other sub-
stance. In the following table, in which the ratios of the absolute
temperatures of the alcohols and acids to those ef water are given, it
is seen that the coefficient (C) of increase or decrease of the ratio per
unit rise of temperature of the standard substance (water) is a
constant.

The true values of C having been obtained graphically, the ratios
found were corrected and the absolute temperatures of the acids and
alcohols at the required pressures were calculated by multiplying those
of water by the corrected ratios, the absolute temperatures were then
reduced to degrees Centigrade and compared with the observed
results.

It will be seen by referring to the annexed tables, that, with one or
two exceptions, the two sets of numbers agree well. Obviously this is
a simple method for smoothing the observed results of vapour-pressure
dleterminations.

773

0080. T OTS80- T 70Z0-T | ¥0Z0-T 166-0 6066-0 . SZP6-O 0116-0
6080. T PO8O- T 6020- T ¢c0z0- T 0266-0 £066- 0 81F6-0 6016-0
O180- T £080 -T Z120- 9020- T £266 -0 1166-0 8TP6- 0 8016-0
ST80- T O1so-T L120. $120-T O0£66- O T166-0 PIP6-0 LO16-0
0280. T 8180-T 1220. | S220-T 66-0 9266-0 Lz¥6-0 cOl6- 0
L280- T ¥2S0- T ra On Z20- T#66- 0 $£66- 0 PEVG6- O 016-0
PESO- Z4€80-T €€20- LE<0- 6766-0 €°66- 0 TEh6- 0 2016-0
TPSO- 0€8s0-T OFzO- L¥Z0- £°66- 0 9966-0 $2P6-0 6606-0
OSso.- 6780: T 0S20- T920- £966 -O LL66-0 O€F6- 0 £606 -0
6°80. 0980. T cco. 1920. 2166-0 €866-0 6246-0 1606-0
0Z80- PlL80-T 2920. ~9Z0- P866-0 0000 -T 6£16-0 0606 -0
880. 0680: T 8220. 2920- L666- 0 L666: 0 CEr6- 0 ¢c206-0
8680. TI160- T 1620. 2620 - O100- T 8100-T T°P6- 0 0806 -O
8160. 8T60-T O1€0- TTE0- 0€00- T 1200: T Scr6-0 9406-0
0c60. ¥960- T PTEO.- PPREO- ~900- T ¥900- T PLEG- O 9906 -O

Ee Oe Boe oe oe
Sees ae ee eee
Sooo0oooooSooSoSsSSO

Re ee ee ee Be Boe oe oe |
cooooooeooSsSO

*$c00-0=0 |. ‘2200 -0=0 |. °GZ00-0=0 |. ' ‘*6000-0=0 |. ‘L000-0=0
-poyefnaye poearosqa -paqeyno|vy poarosqa poarosqg prasesqg

° , P , *‘poarosqgo
poqyenoeg pozenopeVp pozelno[Rg

‘HO*H'O "HO“H‘O "HOH ‘HO*HO

*s[OYOoTV

"WU OGL PUD “ww YG Usemjeq sainssacd-snodDA 4D La}DAA 07 SpLopP pun sjoyoojp Jo saungosadway, aynjosqy ey} fo ovat

=
=)
mM
“A
oO
—
i)
“4
Z~
b—
a
jon}
=
e
=
a
A
io)
DM
=
few}
—
ee
i.)
_
=

$Z0Z-

CZ0Z.-

8$80Z-

0802

$£02- T
$£0Z-T
OFOZG- T
PPOZ- T
6F0Z- T
PC0Z- T
090Z- T
990Z- T
€L0Z- T
Z80Z - I
OOTZ: T

0¢02- T
2802-1
9202: T
Z80Z- I
L802: T
1802-1
6802- T
6802-T
0€0Z- T
6h0Z-1
1¢0z- T
~S0Z- T
€80Z- 1
2802-1
[11Z, T

6EVI-T
6EPFI-T
6EFI-T
6EFT- 1
OFFI-T
OFFI: T
OFFI T
IPPI-T
ZPVI- I
Crrit-T
OFFT- T
CPT: 1
LYFI-I
SPPI-T
OcFI- TI

6EFT- T
Seri. TI
SEPT: 1
CEPI. TI
VVPL-T
: v1: I
oF ¥ L- T
TPPI-T
SEFL-T
PEFL-L
TIFI-T
ZEPT- 1
COVI-T
PORT. T

“pozyeno]eyy

*‘poatosgo

‘2000-00
*poszepnorey |

"pas.tosqg

6891. T
6801- T
680L- T
6801- T
680L- T
S80L-T
SSOI-T
SSOL-T
S80I- T
SHOL- T
Z801-T
L801: T
L801- T

T60T- T
6L01- I
ZS8OT- T
680T- T
9601: T
S60I- T
8601-1
ZSOI- T
TS6L- T
ZLOL- I
6801-1
L801-T
O6OT- T
SLOL-T
€80l-T

P1SO- I
L0¢0- T
S6F0- T
68F0- T
OLVO- T
GYFO- T
Soro: T
LYVt0-T
O¢F0- T |
0ZP0- T |
TOFO- IT |
O8go- T |

9ZC0- T
L0¢0.- T
T6F0- T
E8t0- I
OLFO- T
290: I
LSt0-T
L¥¥0-T
OVPO- T
PEO: I
LOFO.- T
O8¢o- L
9980-T
€880-1
PLEO. T

9°80. T
1280. T
1220. T

ecoo- T
9F00- T
8£00- T
$200- T
L100-T
OF00- T
0666-0
0866:
£966.
966.
6266.
9066.-

€F00-T
€800- T
Lz00-T
LzZ00-T
e100: T
6200: T
L100. T
P666-0
O0L66-0
Cr66- 0
0866-0
L686: 0
8286. O886 -O
Herr 2286-0
vhlé.- 9926

‘1000.0=9 |
*payepno[ey |

‘poalasqg

‘H000"H'O

‘0F00-.0=O

*pozepno[eg ‘poarosqa,

‘HOOO“HO

‘L400-0=0

“poyepno]B,y *‘paarasqg

‘HOOO'H

“sploy

A
—
1S)
a
2)
ar
Z
<4
S
ro)
io)
=)
Zz
=
nM
J
=)
0)
o
ro)
4
<
fs
>)
Nn
ica)
(a=
=)
7
nn
ie)
2
yi
c
:
>

O€ - 621
T9- LeT
09- S21
VS. S21
9T- 161
TZ-SIT
0-9IT
8-ZIT
89-60T
86- SOT
9F - TOT
P9- 96
20-06
LS. 18
€6- 29

$6 -90T
€- SOT
PP- SOT
98- TOT
LI- 66
06- 96
92-6
ce. 16
8&-88
8-8
¢9-08
L0- 94
64-69
94-19
PO- GP

0-LOT
§-COT
€- 801
r- 10L

SP- 96
LL-¥6
00: €6
00- 16
16-88
OL: 98
VL-08
T¢g- 18
FP. SL
90-.¢L
88-04

¢$.99
0¢- 09
L9. ZS
62- OF

QF MIOFNOOANROPOAS

8- LL
€2-9L

ANMDONHOHO

NANFFTOHHDOSCORKRRER

DODFAHOW MM MMON AAS

ef et EN

1Z- 99
£9. 9
Zh. 29
99-09
0-89
LI-9S
06- 8S
£6- OS
OL: LE
62-4
8. 0F
09. ¢¢
89. 62
91-12
8-6

SZ383s3
wD uO ovo

oRRSISSS

SBPSALYAFRQSOOSOMW NN AAD

“pozyelno[ey

‘joyooys [Auwosy

*‘poqe[no[rg

“poezyelno[ey *“‘poatesqg

*‘loyooye [Aynqosy

*poyelno[ey | ‘poatosqg

“pezyepnoley “pealesg a

‘joyooys jAdoag

‘joyoors [AyIaT

*joyoors [AqyoyT

* uIUE
Ul ‘salg

‘pa.vdwuos sainssaid-inody 4 pazvasL0Q) puw paasasqy¢)

wn
<a)
f=
(=)
D
7
i
jon]
mn
om}
=)
ro)
qm
<
>
i
io)
m
az
[o)
_
B
z
P=
(on)
3
B
&
=)
Z
co)
5
=
ies]
Oo
_
fox)

- PLT
*GL1
j- OLT
-L9T
- SOT
3+ GOT
- 6ST
- SST
- TST
- LVT
- SVT
68 -98T
8Z- 621
€¢- 611
T6- FOL

- CLT
7 |
OLT
LOT
¢91
691
-6S1
SST
1st
LYL
a at
9€T
-6261
-611
FOL

WwW OFMWWONSDAO®

0-€ST
- TST
“SPL
“OFT
- SPT
- OFT
“LET
- PET
7 OST
-92T
- 121
“CIT
-80T
- 86

7 £8

DADGOAFGOOM ANDO

96- GET
L0-S8&T
€8-CsI
68- SSI
C8- OST
90-821
C6- F321
OL: 12T
68-LZ11
6L- E11
9L-80T
PP -SOl
P-96
98-98
0-ZL

- LET
- CEL
-€8I
1&1
$Z1
-SEL
Té1
-LIT
-€1T
-8S01
€01
96

-98

€F- 6IT
O¢-9LT
6-811
8€-1IT
89- 80T
OL- SOT
OF- GOL
16. 86
£8 -56
SP -06
L1- $8
€¢.6L
06-12
02-69
0L-9P

TODO Mm ASA ROH
OD
Oo

le)
@

CP.
TP.

[v-
LL.

cL.
19-
og.
ge.
86.
26.
18.
O1- 98
OF- 9F
ws | 18

“peze[no[ey

*‘poadosqg

*‘pozernoyeg | ‘poasesqg

“‘peqelnoyuy | ‘peatosqg

*‘poyepnoreg | “‘pessesqa

“plow O1Le[Baosy

*proe o1A4gnqosy

‘pros o1uordorg

"plow oyooy

“poyetnoieg | ‘peasesqo
|

*‘ploe orms07

‘pouuiyu0o—paundwoo sainssaid-snody 4 pajreL10Q pun paasesq¢Q

LXXIV.—On the Magnetic Rotation of Mixtures of Water with some of
the Acids of the Fatty Series, with Alcohol, and with Sulphuric Acid ;
and Observations on Water of Crystallisation.

By W. H. Peraiy, Ph.D., F.R.S.

From my previous work on the magnetic rotation of compounds, it
will be seen that the molecular rotation of water, which is taken as
unity, is not the same as the sum of the values of oxygen and two of
hydrogen, as deduced from the molecular rotation of other compounds.
Thus, hydrogen is found to be 0254, whilst oxygem im hydroxyl
varies from 0°194 in ordinary alcohols to 0°137 in monobasic acids,
and is 0°261 in carbonyl, so that, taking the lowest numbers, it gives
H, + O = 0°645, and taking the highest it is 0°769, instead of 1-0.
From these facts, it appeared that the determination of the magnetic
rotation of hydrated compounds might give numbers which would
indicate whether they still contained water, or whether the water and
the substances with which it was mixed had combined so as te form
new compounds. Thus, if the former were the case, the numbers
should represent the value of the compound + the value of water; if
the latter, the values should be lower than this. For example, if
formic acid were mixed with water in the proportion of equal mole-
cules of each, we may either get HCOOH + H,O or HC(HQ);. The

first would give—

whilst the latter would give a number considerably lower than this ;
for instance, taking the highest values for H, + O given above, it

would be—

The same thing would be true of acetic acid and other members of

the fatty series.
Since it has been believed that some of the fatty acids do unite
with water to form trihydric alcohols, their examination appeared to

be of special interest.
It is known that on adding water to formic acid, using molecular

VOL. XLIX. 3 6

778 PERKIN: MAGNETIC ROTATION OF MIXTURES OF WATER

proportions, a rise of temperature takes place, with an increase of
density ; the product also has a constant boiling point of 107° (Roscoe,
Chem. Soc. J., 1862, 15, 271), giving it the appearance of a definite
compound.

Acetic acid, on the other hand, when mixed with water—molecular
proportions being used—gives rise to a very considerable reduction of
temperature, but at the same time yields a product of greater density
than acetic acid itself, which is very remarkable. The boiling point
of this product is only approximately constant.

Propionic acid was found to behave in the same way as acetic acid,
a very considerable reduction of temperature taking place on the
addition of water, whilst the density was increased.

Butyric and isobutyric acids were also found to give a reduction of
temperature when diluted with water, using about molecular propor-
tions, but the densities were not taken, nor were they further
examined.

Whilst examining these acids, it was thought advisable to deter-
mine the rotation of some hydrated substance which could not be
a definite compound, for the sake of comparison, and for this purpose
a mixture of absolute alcohol and water in molecular proportions was
used. The following are the particulars respecting the products
examined and their magnetic rotation :—

Hydrated Formic Acid.

The density of the formic acid used was a , 1°22734. It was mixed
5

with water in the proportion of one molecule of acid to one of water ;
this caused a considerable rise of temperature. The density of the
product was—

4
d=, 11829,

a)®” 1-16977,
15

25° 4.
dee 1:16460.

If the water and formic acid had mixed without any change of
volume the density would have been—
iS
15°
therefore condensation had taken place sufficient to cause a variation
of 0°01619 in density.

d—, 115358,

WITH SOME OF THE ACIDS OF THE FATTY SERIES. 779

The measurements of the magnetic rotation gave the following
numbers :—

Sp. rotation. Mol. rotation.
0°8767 2°671
0°8726 2°659
0°8752 2°667

0°8748 2°666

Hydrated Acetic Acid.

The acetic acid used had a density of a, 1:05704. It was diluted

in the proportion of one molecule of acid to one molecule of water.
Considerable reduction of temperature took place on making the mix-
ture. The density of the product was—

15°
15°

20° 107163,
20

d=”. 107566,

d=”. 1-06786.
25
The density of the above mixture if no change of volume had taken
place would be—

on 1:04331.

1 °
Condensation had therefore occurred—although a large reduction
of temperature had taken place when making the mixture—equal
to a change of density of 0°03235, or twice as much as that in the
case of formic acid.
The numbers obtained for the magnetic rotation were as fol-
lows :—
t. Sp. rotation. Mol. rotation.
16:0° 0°8818 3°555
16:0 0°8817 3°555
19°4 0°8800 3°558
19°4 0°8784 3551
19°4 0°8784 3°551

180° 0°8801 3°554

780 PERKIN: MAGNETIC ROTATION OF MIXTURES OF WATER

Hydrated Propionic Acid.

The density of the acid used was a. 0°99833. It was mixed in the
proportion of equal molecules. Considerable reduction of temperature
took place during the operation.

The density of the product was as follows :—
£
4°
15°
15°

d>- 1-01729.

25°

d-=_ 1:08448,

d-?_ 1-02494,

If no change of volume had occurred, it should have been
in 099863. Therefore there was an increase of density equal to

ov
0-02631.

The determination of the magnetic rotation gave the following
numbers :—

t. Sp. rotation. Mol. rotation.
19°2° 09034 4519
19-2 0°8964 4485
19°2 0°8996 4°500
19°5 0°9030 4519
19°6 0°9027 4°525
19:1 0°9013 4516
19°1 09022 4°520
19°4° 0°9012 4512

Hydrated Ethyl Alcohol.

The alcohol and water were mixed in molecular proportions, as in
the other cases. As is well known, a result of this operation is a
rise of temperature. The density of the product was—

d a 087682,

15°
d—
15°

086862,

25° ~
d 0°86178.

€ °o
25

WITH SOME OF THE ACIDS OF THE FATTY SERIES. 78i

If no change of volume had taken place the, density would have been
os , 0°85837, showing that condensation had taken place to the extent
»

of 0°01025 increase in density.
The following numbers were obtained for its magnetic rotation :—

t. Sp. rotation. Mol. rotation.
19°4° 0°9206 3°782
19°4 0°9226 3°790
19°4 0°9224 3°789
19°4° 0°9219 3°787

If now a comparison of these results be made with the calculated
values of the products as composed of water and the acids or
alcohols, we get the following. :—

Calculated. —
ae ee 2606
HO reo 832 3364
a. oe 4512
BO 1900 f 3720 3787

From these numbers it will be: seen that these products behave as
composed of the acids or alcohol and water, and not, as was considered
to be the case, with the acids as trihydrie alcohols, which would have
given lower molecular rotations, in fact in the case of acetic and pro-
pionic acids, the numbers are inelined to be a little high. This, how-
ever, may be accidental.

The molecular rotation of glycerol, which has the same composition
as propionic acid + one molecule of water, is 4111. The distribution
of the hydroxyl in this, however, is not the same as it would be in a
compound derived from water and propionic acid, if formed.

The foregoing results are further evidence of the fact that although
thermal changes and alterations of density result from chemical action,
yet that they themselves are not evidence of chemical action. Bussy
and Buignet have shown, for example, that when equal weights of
chloroform and ethyl oxide are mixed, that a rise of temperature of
14°4° takes place, whereas chloroform and alcohol only give a rise of
2°90°, and chloroform and carbon bisulphide give a reduction of 5°
(Compt. rend., 59, 676). I have also lately tried experiments roughly

782 PERKIN: MAGNETIC ROTATION OF MIXTURES OF WATER

with the following mixtures, which all give a reduction of tempera-
ture, viz. :—

Heptane and alcohol.
Tetrachloride of carbon and alcohol.
- ™ and acetic acid.
Ethy] iodide and alcohol.
Carbon bisulphide and acetic acid.

Tetrachloride of carbon with ether gives a rise of temperature, but
not nearly so much:as chloroform. In these cases, we have no ground
for supposing that.any chemical change results from the mixing of
the products.

After obtaining the foregoing results with the fatty acids and
alcohol, it was thought that it would be interesting to examine sul-
phuric acid in its pure and hydrated condition, as it is generally sup-
posed to form two or more well-defined compounds with water.

The first might be either H.SO, + H,O or (HO),SO. The pro-
duct of this composition solidifies at 8° or 9°, forming six-sided
prisms, and is known as glacial sulphuric acid.

The second might be either H,SO, + 2H,0 or (HO),SO +H.0,
or (HO),S. Sulphuric acid mixed with this proportion of water
gives a preduct with the maximum condensation which results from
the union ofthis acid with water.

The third might be either H,SO, + 3H,0 or one of the preceding
compounds plus water. It must-contain water. These three have all
been examined.

As most of my work on the magnetic rotation has been in reference
to compounds containing carbon, it was thought advisable to examine
not only sulphuric acid, but also one of its ethers, to see if the relation-
ship of the acid to the ethers would be at all comparable with that
existing between an acid containing carbon and its ether. This would
serve as an indication as to whether sulphuric acid behaves in re-
lation to its magnetic rotation in a way analogous to other substances
which have hitherto been examined. Methyl sulphate was selected for

this purpose.

Sulphuric Acid, H,SQ,.

For the acid used, I am indebted to the kindness of Dr. Messel. It
was made by the admixture of fuming acid to acid of the ordinary
strength, the percentage being determined both by volumetric and
gravimetric methods. The density determinations gave—

15° ,,
dy=. 183904,

ES AY a a ae mite a dem

F ad

2
2

d—, 183265,

2
15°

At 7 this will be 1°83748. Lunge (Jour. Soc. Chem. Ind., 2,

0°
5°
5°

1°83562,

WITH SOME OF THE ACIDS OF THE FATTY SERIES.

1883, 279) found for acid of this quality - 1:8384.

The determinations of the magnetic rotation gave—

t. Sp. rotation. Mol. rotation.
19°5° 0°7809 2316
20°0 0°7809 2°316
20°5 0°7810 2°317
21°0 0°7787 2°310
20°25° 0°7804 2°315

Hydrated Sulphuric Acid, H,SO, + H,0.

The density determinations of this product gave the following

numbers :—

a 5°

15°

t This at

tion :—
i t. Sp. rotation.
} 20°0° 0°8799
20°0 0°8764
20°0 0°8773
20°4 0°8769
20°4 0°8770

15°
20°

d—, 1:77423,
2U

25"
25°

177806,

1:77071.

is 1°77655, and contains 844 per cent. H,SO,. The density

: of acid containing 84 per cent. H,SO, is given as 1°7770.
The following numbers were obtained for its magnetic rota-

Mol. rotation.

3°196
3°184
3°187
3°186
3°186

20°16°

0°8775

3188

784 PERKIN: MAGNETIC ROTATION OF MIXTURES OF WATER

Hydrated Sulphuric Acid, H,SO, + 2H,0.
The density determinations of this were as follows :—

a!®” 1-65084,
15

a2" 1-64754,
20

d=. 1-64467.
25

The numbers obtained for its magnetic rotation were—
t. Sp. rotation. Mol. rotation.

20°5° 0°9101 4113

20°5 09099 4112

21-0 09109 4117

21:0 0°9091 4-109

20°75° 0°9100 4113

Hydrated Sulphuric Acid, H,SO; + 3H,0.
The density of this substance was found to be—

a)” 1.55064,
15

.
ds 154754,
d=”. 1-54493.

The magnetic rotation determinations gave—

t. Sp. rotation. Mol. rotation.
23° 0:9270 5°064
23: 0°9245 5°050
23 0°9291 5°075
23 0°9281 5°069

23° 09272 5°064

Methyl Sulphate.

This ether was prepared by treating methyl alcohol with four or
five times its weight of 60 per cent. fuming sulphuric acid, distilling
the mixture, washing the ether first with water and then with dilute

WITH SOME OF THE ACIDS OF THE FATTY SERIES.

carbonate of sodium solution, quickly drying over potassium carbo-
nate, and distilling several times. As thus obtained, it boiled between
188'°3—188°6° (corr.). The density determinations gave—

di® 1-33344,
15

20" 132757,
20

25°
gR0

«ov

d
d 1:32386.

The magnetic rotations obtained were as follows :—

t. Sp. rotation. Mol. rotation.
22°8° 0°7590 4°007
22°8 0°7605 4015
22°8 0°7611 4018
22°8° 0°7602 4°013

It will be best perhaps first to consider the rotation of sulphuric

acid in relation to methyl sulphate.

If we subtract the rotation of the acid from the ether, we get the

following result :—

Methyl sulphate .......... 40130
Sulphuric acid............ 2°3150
1:6980

This is the influence of replacing two of hydrogen by two of methyl,

or 0°849 for one of methyl.

Now if we take some of the fatty acids and their ethers, and sub-

tract them, we get—

Methyl formate............ 2°495
ee 1671

0824
Methyl acetate ............ 3°362
Bas 06 46 0064 eadbes 2°525

0°837
Methyl butyrate ........... 5°387

Dutyrie O61 20 cc cccccccece

786 PERKIN: MAGNETIC ROTATION OF MIXTURES OF WATER

It will be seen that the number obtained from sulphate of methyl
agrees closely with that of methyl formate and methyl acetate, but
not so well with methyl butyrate, which is the second member of the
true homologous series in which this last number for methyl is con-
stant. It is, however, I think, quite natural that owing to its highly
chlorous nature, it should be most nearly comparable with formic or
acetic compounds, and there is, therefore, no reason to doubt that sul-
phuric acid and its compounds, in regard to their magnetic rotation,
behave in a manner analogous to other compounds.

With respect to the rotations of the hydrated sulphuric acids, the
results are not so clear as in the case of the hydrated fatty acids, as
the following comparisons will show :—

Calculated. Found.
Sulphuric acid .... samt '
HO 1-000 3°315 3188
Sulphuric acid .... aaet ,
9-000 4°315 4113
2°315

rhe \ 5315 5-064

Here we get lower rotations indicating combination. In the first
instance, we get an increase of rotation, owing to the water added, of
0°873, in the second of 0°925, and in the third of 0°951, instead of
1:0, which is the value of water, the last number being, however, very
near to this. These numbers indicate that the largest amount of com-
bination takes place when the first molecule is added, and but very
little when the third is added, in fact it would appear that sulphuric
acid with one molecule of water forms the compound (HO),SO, but
that all the acid cannot exist as this without the presence of a larger
amount of water. Now if we take the rotation of the product made
by adding three molecules of water to sulphuric acid, and subtract
the value of sulphuric acid from it, we get—

Influence of 3 mols. °
of water added . \ 5740

This value is equal to that of two molecules of water + 0°749;
therefore if the compound (HO),SO has been formed its rotation
will be that of sulphuric acid + this number. Is this likely to be the
case ?

I have lately been examining the rotation of the anhydrides of the

WITH SOME OF THE ACIDS OF THE FATTY SERIES. 787

fatty acids, and the results obtained I think will go to answer this
question, seeing that the rotation of the hydrated anhydrides or acids
are also known. There is a close analogy between the formation of an
acid by the action of water on an anhydride of a fatty acid and the
formation of such a compound as (HO),SO from sulphuric acid.
Thus, taking acetic anhydride—

in both cases one atom of oxygen of the compound and one mole-
cule of water go to form two hydroxyls; in the one case the product
splits up into two molecules, in the other it remains a single one.

The molecular rotations of acetic and propionic anhydrides are given

below, and subtracted from the molecular rotation of the two mole-
cules of the respective acids which result from their combination
with water. This will give a number representing the increase of
rotation due to the combination of a molecule of water.

Rotation of acetic acid = 2°525 x 2 = 5°050
acetic anhydride

”

Rotation of propionic acid=3'462 x 2 = 6°924
propionic anhydride .... 6185

0°739

Now it will be seen that these numbers correspond very closely with
the residual number obtained above in the case of sulphuric acid, viz.,
0°749, so that the probability of the compound (HO),SO being the re-
sult of the addition of water to sulphuric acid becomes very strong,
especially as the rotations cannot in any way be made to agree with
the formation of (HO),S, the numbers being too high; in fact only
one chemical compound seems to result from the action of water on
sulphuric acid, namely, (HO),SO.

Experiments are in progress on the other sulphuric compounds, and
it is also proposed to examine sulphurous compounds.

788 PERKIN: MAGNETIC ROTATION OF MIXTURES OF WATER

Whilst working on hydrated products, my attention has been
drawn to the subject of water of crystallisation. The point
which has struck me most is the want of consistency as to the
presence or absence of water of crystallisation in the simple salts of
metals belonging to the same class. Take, for example, the salts of
silver, potassium, and sodium. The silver salts are mostly anhydrous,
the larger number of potassium salts are also anhydrous, whilst
sodium salts generally contain water of crystallisation. Moreover,
when analogous salts of similar metals contain water of crystallisation,
the proportions are not the same, as in K,CQ,, and Na,CO;. These
instances might be greatly multiplied. Amongst carbon compounds
there is also a very remarkable instance of this inconsistency. I refer
to the crystalline compound of methyl bromide and water (from my
determinations, this consists of 1 mol. methyl bromide to about 10 mols.
H,0; Merritt gives it as probably containing 20 mols. H,O, J. pr.
Chem. [2], 18, 293). This substance stands quite alone, no other
related halogen carbon compound behaving in a similar manner.

These inconsistencies are in fact so great that it is impossible to
believe that water of crystallisation has any relationship to chemical
combination. Of course hydrogen and oxygen may be contained in
some salts in the proportions found in water, and be chemically com-
bined,* but then they are not in the form of water. For example,
copper and magnesium sulphates do not lose what is supposed to be
the last molecule of water which is contained in the crystallised salt
unless very strongly heated. This hydrogen and oxygen, however,
may not be in the form of water, but the compounds may correspond
to the substance which is obtained by mixing sulphuric acid with
water, viz., (HO),SO, and be constituted thus—

R’0,
HO}, f 80-

What, then, can be the relationship of water of crystallisation to
chemical compounds? It appears to me that it is in all probability
purely physical, its presence with the compound being necessary for
the building up of the crystalline form which can most readily be |
produced.

We know that the crystalline form of a substance containing water
of crystallisation is essentially connected with the amount of water it
contains. For example, sodium carbonate may be taken, which crys-
tallises with one, five, six, eight, and ten molecules of water, varying

* By chemical combination, I mean that which takes place between molecules of
different kinds, whereby they are more or less broken up, and the constituents re-
arranged so as to form new molecules; or that which takes place when different
molecules act on each other so as to produce new molecules not containing either
of the original ones used in their production.

WITH SOME OF THE ACIDS OF THE FATTY SERIES. 789 —

in form (that with 10 mols. being the most easily produced, though,
if chemical combination had to do with its formation, we should have
expected it to be the most difficult, as it should be the least stable),

Again in products where the proportion of water is constant, the
crystalline form often remains practically unchanged although the
other constituents may be varied. Take, for example, the large
number of alums which contain 24 molecules, or nearly half their
weight of water; also the large number of substances isomorphous
with ammonium magnesium sulphate containing 6 mols. H,O, and
the sulphates of magnesium, nickel, and zinc. The sodium salts, with
10 mols. H,0, such as the carbonate, sulphate, chromate, and pyro-
phosphate, are also all said to belong to the oblique prismatic system,
showing the important part water must play in reference to form.

The amount of water of crystallisation in some of the above-men-
tioned substances, if taken in reference to the amount of water in the
salts from which they are formed, is also worth considering. Take
potassium and sodium alums. The former is made from potassium
sulphate, which is anhydrous, and aluminium sulphate, which contains
18 mols. H,O; the alum contains 6 mols. more H,O. The latter is made
from sodium sulphate, which usually contains 10 mols. H,0, and sul-
phate of alumina, which contains 18, making 28 in all; the alum con-
tains 4 mols. less. Again, take iron alum. The potassium sulphate is
anhydrous ; the ferric sulphate contains 9 mols. H,O. This is 15 mols.
less than is contained in the resulting alam. With the compounds of
the ammonium magnesium sulphate class we find much the same thing.
Magnesium sulphate contains 7 mols. H,0; the so-called double salt
contains 6, or one less. When copper sulphate, which usually con-
tains only 5 mols. H,0, is used in place of magnesium sulphate, the
so-called double salt also contains 6 mols., or in this case one more
than the copper sulphate used.

From these changes in the amount of water of crystallisation, it
seems that we cannot assume that its relation to a salt or other com-
pound has any connection with chemical combination, otherwise the
influence of the water contained in the substances used in the prepa-
ration of these compounds would be in some way exhibited in the
product formed. For example, sodium salts nearly always contain a
large amount of water of crystallisation. If this were due to chemical
union, this should certainly manifest itself in sodium alum, but it does
not. We find, however, in the isomorphous compounds referred to,
that the crystalline form and the proportion of water of crystal-
lisation remain the same in each group, and if, therefore, the water
is not chemically combined, its association with the chemical com-
pound would appear to be in relation to the building up of the crys-
talline form. It is difficult to see what other part it can play.

790 RAMSAY AND YOUNG: EVAPORATION

Why, then, do some compounds crystallise without, and some with
water of crystallisation ? The answer to this is, probably, that if a salt
by itself can produce the most easily formed crystalline structure, that
it will naturally do so, and be anhydrous, but if, on the other hand, it
cannot do so, that it will associate itself with that number of molecules
of water with which it can most readily produce the crystalline form.

These observations would equally apply to compounds crystallising
with alcohol, acetic acid, benzene, &c., and to some double salts, one
or more of the constituents acting like water of crystallisation.

There is one point of interest which should result if the above view
respecting water of crystallisation be correct, and that is in reference
to the subject of solution. It is evident that if water of crystallisa-
tion be only related to crystalline form, this water will no longer
be attached to the compound with which it is associated on the
breaking up of this form by solution. And this is believed to be the
case by many who have studied the subject of solution.

I have ventured to bring these observations forward, although in a
rather incomplete way, thinking that the subject of water of crystalli-
sation is well worth considering from this point of view.

LXXV.—Evaporation and Dissociation. Part IV. <A Study
of the Thermal Properties of Acetic Acid.

By Wittiam Ramsay, Ph.D., and Srpyey Youne, D.Sc.

1. The abnormal behaviour of acetic acid and homologous acids has
rendered them the subject of numerous investigations. At mode-
rately low temperatures, these substances invariably show vapour-
densities, which would lead to the conclusion that the structure of their
molecules is not so simple as that expressed by their formule. But
although many experiments have been made, especially with acetic
acid, no complete research is on record, showing the influence of
pressure and temperature on the density of its vapour, and on the heats
of volatilisation. It is only necessary to name Playfair and Wanklyn,
Cahours, Bineau, Horstmann, and Naumann, to indicate the import-
ance attached to this subject. The results of these experimenters
shall be considered later.

2. Material used in this Research.—The sample of acid used was a
portion of the stock with which we have already experimented. To
quote from a previous paper on the vapour-pressures of acetic acid

AND DISSOCIATION. 791

(Trans. Roy. Soc., 1884, ii, 469), “The acid which we used for these
experiments was a portion of a stock of glacial acid obtained for
laboratory purposes several years ago. The liquid portions had been
poured off from time to time, as required for laboratory use. A very
complete series of fractionations has thus unwittingly been carried out,
and it is now so pure that at a temperature slightly below its melting
point, it is completely solid.” After an aqueous portion of about
400 c.c. had distilled off, the boiling point became practically constant,
and remained constant until the last trace of the remaining 200 c.c.
had distilled over. The rise in temperature did not amount to 0°05°
during the distillation of the last portion. It will be afterwards
shown that the density of this sample is nearly the same as that
found by Perkin, who took especial pains to purify and dry his acid;
but we are of opinion that any attempt to dry the acid by the ordinary
desiccatory reagents introduces acetic anhydride. It is doubtful
whether the want of constancy in the boiling point of Perkin’s sample
of acid was due to this cause, or to the presence of homologous acids ;
but it may here be noticed that we ourselves attempted to obtain a
product of constant boiling point from a specimen of ordinary glacial
acid, freshly purchased, and were unable to doso. As a test of purity,
we may mention that identical vapour-pressures were obtained with
several samples of acid fractionated from the stock above referred to,
by both the dynamical and statical methods, and this is, according to
Regnault, a crucial test of purity. He says :—“ Lorsqu’une liquide
renferme une portion méme extrémement petite d’une autre substance
volatile, les deux méthodes donnent des valeurs différentes pour les
forces élastiques de sa vapeur 4 la méme température; et c’est un
moyen extrémement délicat pour juger de l’homogénéité d’une sub-
stance volatile.” (Mémoires de ’ Academie, 26, 341.)

3. Apparatus Employed.--As a full description of the apparatus
used has been given ina paper on the thermal constants of ether,
recently communicated to the Royal Society, it is unnecessary to
describe them further here, than merely to state that the vapour-
densities at low temperatures were determined by a modification of
Hofmann’s apparatus, in which pressure, temperature, and volume
could be altered at will; and at high temperatures by a modified
Andrews’ apparatus, provided with two air-gauges, one for registering
high and the other low pressures. The calibration of the gauges
and of the volume-tube has, moreover, been fully described, and also
the method of reducing the results. The temperatures were controlled
by the method which we have already described in the Transactions
of the Chemical Society, 1885, 640. The results recently obtained by
Amagat for the compressibility of air were introduced in calculating
pressures (Compt. rend., 99, 1153).

792 RAMSAY AND YOUNG: EVAPORATION

Experimental Results.

4. Vapour-pressures at Low Temperatures.—These constants were
determined by us in 1884. A full account of the method of research
and a statement of the results is to be found in the Phil. Trans., 1884,
465, and in this Journal (Trans., 1885, 42).

5. Vapour-densities in Hofmann’s Apparatus.—The acetic acid
taken was weighed out in a small tube, and introduced into the
Hofmann’s apparatus. We append an example of the method of
calculation, which will give an idea of the corrections introduced.

Barometric pressure 75983 mm.
Height of mercury in gauge 536°75 _,,
Modified atmospheric pressure 223°05_,,
Height of mercury in Hofmann’stube.. 209°9
Height of mercury in bottle

Difference

Correction for temperature of mercury in

Vapour-pressure of mercury

Corrected height of mercury column at
temperature of room

Correction for capillarity of differential

Uncorrected pressure of vapour, 222°95—

209°55 =
Corrected to 0° 13°35 _,,
Volume of gas at temperature 50° 170°06 c.c.
Weight 0°01126 gram.
Vapour-density

6. Weight taken, 0°01126 gram.
Pressure of Alcohol Vapour 220 mm.; Temp. 50°.

Vapour-

Vapour-
density.

density. Pressure. Volume.

Volume.

170-06 c.c. 49 ‘58 79°96 c.c. 53°27
150°10 ,, 48 ‘88 69 *97 53-83
134°88 ,, 49°43 60 00 53°84
119-80 ,, 51°83 40°17 54°79
105 ‘00 51°25 33 *2 60 -78
(liquid
present)
90 ‘08 51°94 . 25 °3 78-08
(liquid
present)

AND DISSOCIATION.

7. Pressure of Alcohol Vapour 761 mm.; Temp. 78°4°.

Series I.

Vapour-

density. Pressure.

Pressure. Volume.

Vapour-
density.

39°91 125°3 mm.
39°14 141°3_—Ssé=«;
38 °97 164°2_ ,,
40°56 i aa
42°61
42°74 194 °2 Pm
44°57
46 *40 197 °3 -
47°97

49°17
50 °87
50°93
53 °64
(liquid ?)
68°81
(liquid)
80 92
(liquid)

Series IT.

Vapour-

density. Pressure. Volume.

Pressure.

Vapour-
density.

39° ‘ 50°01
39°
40°
40°
41°
41°
41°
42°
42° €
42°
43°

38° 59°97 c.c.

(liquid)
57°38
53°43

8. Weight = 0:03565.

Temperature the same.

Vapour-

density. Pressure.

Pressure. Volume.

Vapour-
density.

137°8 c.c. 44°65 | 188°55 mm.
76°0 ,, 47°99
651 ,, 49°16 | 190°55_,,
49:0 ,, 50°31
41°88 ,, 51°36 | 192-05 _,,

52 94
(liquid)
55°99
(liquid)
70°38
(liquid)

VOL. XLIX.

RAMSAY AND YOUNG: EVAPORATION

9. Weight = 0°01126.
Jacketed with Acetic Acid; Pressure 314 mm.; Temp. 92°0?.

Vapour

Vapour-
density.

: Pressure. Volume.
density.

Pressure, Volume.

165 °0 c.e. "85 44 *22
135-0 8 “bE | 45°66
105 ‘0 *§ 47°21
70°35
(liquid)

10. Weight = 0°03565.

Temperature the same.

Vapour-

Vapour-
density.

Pressure. Volume. : Pressure. | Volume.
density.

104° 87°0 ,, | 70
135° 65°05,, | 78 7
170° 50°0 ,, 35
234° 35°0 ,, 18
268° 30°05. ,, 13

75°65 mm.| 125°3 e.c. 2-66 7 5 mm.

292°9 i, |
| (liquid ?)

11. Weight = 0°01126.
Pressure of Acetic Acid Vapour 498 mm.; Temp. 105°1°.

Vapour-

Vapour-
density.

Pressure. Volume. :
density.

Pressure. Volume. |

36 06 44-09
37 °9 46°35
39 28 | | 70°31
40°14 | (liquid)
41-89 |

|
35°49 " | 42 *52

AND DISSOCIATION.

12. Weight = 0-03565.

Temperature the same.

Vapour-

" Pressure. Volume.
density.

Pressure. | Volume. density.

39°22 422°4 mm. 20 °0 c.c.
40°63 460°35 a.
42°85 480 °5 16°0 ,,
44°11
45 °02 474°5 7 6
46°76

|
| Vapour-
|

(liquid)

13. Weight = 0°01126.
Pressure of Acetic Acid 751°8 mm.; Temp. 118°2°.

Vapour-
density.

Vapour-

: Pressure. Volume.
density.

Pressure. Volume.

mm. : 36°27
37 °53
39°55
40 *62
43°48
45°02
A5 “89

24°1 mm.| 170° ; 33°39
29 - 33°58
33 93
34°59
35°35
35°93
36 °06

ADOonwnan

14. Weight = 0°03565 gram.
Pressure of Acetic Acid 745°8 mm. ; Temp. 117°95°.

| |
Vapour-
density.

Vapour-

; Pressure. Volume.
ensity.

}
Pressure. Volume.
|

75° ‘ bi 5°73 6768 mm.| 13°0 c.c. 49°26
85° 38: 3°77 7224 , | 101 ,, 59°42
163 *5 :: | 78 | (liquid)
210: “iy oy ae 11°05 55°04
256° 39 °9: 2° (liquid)
328 -6 | 43° 706°5 ,, | 12-0 51°13
386 * 25° ; (liquid)
467 - i 733°5 ,, 8:0 73°93
601° 5° 8° (liquid)
632°

15. Pressure of Aniline 160 mm.; Temp. 132°9°.

RAMSAY AND YOUNG: EVAPORATION

Vapour- , Vapour-
Pressure. Volume. density. Pressure. Volume. density.
101°7 mm.| 128°0 e.c. 34°54 360°4mm.| 30°35 c.c. 41°11
132°7 ,, 95-0 ,, 35 “67 470°7 ,, | 220 ,, 43 42
174°5 ,, 70°2 ,, 36°71 700°2 ,, 13°95 46 °03
214-9 ,. 65°1 ,, 37°98 747°6 . | 13-0 ,. 46°27
285°2 ,, 40°0 ,, 39-43
16. Pressure of Aniline 262°5 mm.; Temp. 147°6°.
- .
Pressure. Volume. ium. Pressure. Volume. aus. )
88°5mm.| 161°0 ce. 32°7 341°8mm.| 36°0 e.c 37°87
133°9 ,, | 100-0 ,, 34°8 427°9 ,, | 28°0 ,, 38°88
181°6 ,, 72°9 ,, 35°19 507°0 ,, | 23°0 ,, 39°95
2251 ,, 580 ,, 35 “69 6025 ,, | 19°0 ,, 40°70
263°5 ,, 482 ,, 36 68 729-7 ,, | 150 ,, 42°57
17. Pressure of Aniline 416 mm.; Temp. 162°5°.
Pressure Volume Vapour- Pressure Volume Vapour-
. : density. ' , density.
95°9mm.| 159-0 cc. | 31°64 | 450°5mm.| 30-0 cc.| 35°71
134°3 ,, | 112-0 ,, 32°07 526-7 ,, 25-0 ,, 36°65
1838 ,, 80:0 ., 32°81 640°3 20°0 ,, 37 °68
274°0 ,, 520 ,, 33°87 729°5 ,, 170 ,, 38-91
373°1 ,, 37°0 ,, 34°95
18. Pressure of Aniline 754-7 mm.; Temp. 184°1°.
Pressure. Volume. —— Pressure. Volume. Vapour-
ensity. density.
98°05mm.| 165°9 c.c. 31°13 425°7mm.| 36°0 c.c 33°05
140-7 ,, | 1149 ,, 31°33 520°4 ,, | 29°0 , 33°56
1882 ,, 85:0 ,, 31°66 6664 , | 22-0 , 34°54 :
242°9 ,, 65°0 ,, 32°07 791°6 ,, 18°0 ,, 35 54
336°8 _s,, 46°0 ,, 32°69

19. For these experiments, four different quantities of acetic acid
were used.

Constants with Modified Andrews’ Apparatus.

With the first and second, the expansion was determined ;

AND DISSOCIATION. 797

with the third, the vapour-pressures and the vapour-densities near
saturation point at the highest temperatures; and with the fourth,
the densities of the saturated and unsaturated vapour.

20. Weights of Substance Employed—tThe weights of the first three
portions were determined by reading the volume at a known tempera-
ture and low pressure, and by determining the sp. gr. of the acid at a
closely approximate temperature. The details are as follows :—

The determination was made in a Sprengel’s tube of the form
recommended by Perkin (Trans., 1884, 443).

Weight of acetic acid at 13°11°............ 39°5025 grams.
Weight of water at 13°20° ................ 37°3558_—,,
Specific gravity at 13°11° (water at 4° = 1:0) 1:05682
VORP DOR cc cecnsctccucéessccesss 0°94622 c.c.
| (These weighings were reduced to a vacuum.)
; Portion I.— Volume at 17°95°.............45. 0°65829 c.c.
Volume at 13°11°, graphically ob-
tained from expansion ........ 0°65593 _,,
Therefore weight .............. 0°69322 gram.
The limit of accuracy in reading was here 1 in 2000.
Portion II.—Volume at 19°77° .............. 0°20456 c.c.
Volume at 16°10° .............. 0°20435_,,
Volume at 13°11°, graphically ob-
SE ciciscnvaansideunacen en 0°20349 ,,
The weight is therefore.......... 021505 gram.
The limit of accuracy in reading was here 1 in 500.
Portion III.—Volume at 14°5° .............. 0°06987 c.c.
Specific gravity at14°5°, calculated 1°0553
Hence weight ............002. 0°07374 gram.

Limit of accuracy 1 in 200.

This weight was used in calculating the vapour-density, but as a
check on its correctness, a measurement at 279°85° was made.

pices aaa nied

Volume at 279°85°............ 0°11223 c.c.
Specific gravity .............. 0°6595
WOMEN 6 cecceseesesceecnesies 0°07401 gram.

This, it will be seen, is within the limits of error.

Portion IV.—The weight of the fourth portion was determined by
comparisons of the volumes of its unsaturated vapour with those of
the third portion at the same pressures. The comparison was made
+ graphically, and from it the weight was found to be equal to
0°005358 gram.

21. First Portion.

RAMSAY AND YOUNG :. EVAPORATION

Expansion of Liquid.

Weight = 0°69322 gram.

|
’ Tempera- Vol..of 1 gram.
Jacketing whee ome on | Vol. of ati |
reduced | .. ss Sp. gr
vapours. to o°, | air ther-| liquid. —
mometer. Read. |Smoothed.
|
mm c.c.
—_ 0° — _ 0°9348 | 1-0697
_ 10 — — | 0-9440 | 1-0593
~- 17°95 | 0°65829 | 0°94962 | 0°9512 1°0513
—_— 20°0 — _— 0 9532 1°0491
_— 30°0 _— 0 -9623 1°0392
BGR cc cece cscs] MOT 40 °0 0°67353 | 0°97160 | 0°9724 1 0284
~— ‘¢6éeccce0sPauee 50°0 0°68099 | 0°98236 | 0:9828 1°0175
9 «0 a cccese] OOO eS 60°0 0°68916 | 0°99415 | 0°9940 1-0060
ape rerer aie 541°2 | 70°0 | 0°69732 | 1°0059 | 1-0052 | 0-9948
Chlorobenzene .. 144°8 80°0 0°70528 | 1°0174 1-0168 0 9835
- ..++| 208°35 | 90-0 | 071307 | 1:0286 | 1-0290 | 0-9718
= ee 292°75 | 100-0 0°72230 | 1°0420 1°0418 0 9599
~. 402 °55 | 110°0 0°73151 | 1°0552 1°0545 0 °9483
~ ...-| 542°8 | 120°0 0°74093 | 1°0688 1 -0682 00-9362
- eee 718 °95 130°0 0°75105 | 1°0834 1°0828 0 °9235
Second Portion. Weight = 0°21505 gram.
Tempera- Vol. of 1 gram.
Jacketin Pressure inn on Vol. of ’ a
8 | reduced | .. ene" Sp. gr.
vapours. | too? | ait ther- | liquid. |
mometer. | Read. |Smoothed.
|
| mm. c.c, |
Chlorobenzene .. .. | 542°8 120° | 0-22991 | 1°0691 | 1-0682 | 0-9362
”» ee: 130 0°23215 1°0795 1°0828 | 0°9235
Bromobenzene....| 372 °65 130 0°23310 1°0839 1°0828 | 0°9235
* ee+-| 495°80 140 0 *23598 1°0973 1 0980 0°9108
a ...»| 649°05 | 150 | 0°23954| 1:1139| 1°1138 | 0-8978
Aniline ..4...00-+| 283 -7 150 0°23954 1°1139 1°1138 0°8978
mw seccecesse| Gee | 160 0°24343 1°1320 1°1301 | 0°8849
» eeeeeeeees| 815°6 | 170 | 0-24666 | 1-1470| 11481 | 0-8710
— . seSenecnce! Gee am 180 0 *25090 1°1667 1°1678 | 0°8563
Methyl salicylate. . | 249°35 | 180 0°25090 | 1°1667 | 1°1678 | 0°8412
i » «e| 880°85 | 190 | 0°25612 | 1°1910| 1°1888 | 0-8412
. » «| 432°35 | 200 | 0-26082 | 1-2128 | 1-2117 | 0-8253
“ -- | 557 °5 210 0°26597 | 1:°2368 1°2360 _ 0°8091
. » ve 710° 220 | 0-27147| 1°2623/ 1°2632 | 0-7916
Bromonaphthalene 235 95 230 | 0°27782 | 1°2919| 1°2933 | 0°7732
- | 303 °35 240 | 0°28455 | 1°3232 1°3270 | 0°7536
= | 386°35 250 0°29423 | 1°3682 1°3640 | 0°7331
a 487 *35 260 0°30235 | 1°4060/| 1-°4060 | 0-7112
ne | 608°75 | 270 | 0°31329 | 1°4569/ 1°4561 | 0-6868
- 751°4 279°9 | 0°32609 | 1°5164 1°5172 | 0°6591
| (280°)

22. The vapour-pressures and vapour-densities were determined
with portions 3 and 4.

Vapour-pressures and Vapour-densities.

a ee ae

eS er me

AND DISSOCIATION.

Portion 4. Weight = 0°005358 gram.

Pressure | Temperature} vy
Jacketing vapours. reduced on air er Mean.
to 0°. thermometer.| Pressure:
mm.
4 Aniline ....00cccccccoes 144°7 130° 1035
' 10140 1040
= |
9” . 204°6 140 1376 )
1387
1384 f 1381
1376
PT ETT TTT T Terr 283 °7 150 1821
1834 |
1830 1847
| 1843
1909 J
9 386°0 | 160 2375
2368 ”
| | 2343 =
| | 2398
| |
Pressure! Tempera- 7
Jacketing vapour. | reduced | ture on | Volume. | Pressure.| P.V. nal
! too’. | 2 ther- | density.
| mometer.
|
mm. c.c. mm. |
Aniline ..........} 416°0 | 162°5° | 1°1630 1413 | 1644 44°12
99 te eee ween! - , 1°0572 1526 1613 44°94
~~ scessocece! ; 0°98623.| 1616 | 1594 45 °49
- |» mm 0°87854 | 1777 | 1561 46 *44
- 99 ” 0°80660 | 1907 | 1538 47°15
, ; ” 0°73540 | 2049 | 1507 48°12
» saene eccee ” » 0 69991 2128 | 1490 48 68
» eoeees re , 0 °66434 | 2212 | 1470 49 *34
, oc ccee 9» ‘ 0°62883 | 2291 | 1441 50°33
cation ma “i 0:59317 | 2390 | 1418 51°16
9 ene % 99 0°55760 | 2456 1370 52°96
| trace of
2496 | liquid
2541
= sebeseaces . | - — 95,46 | Mean 2527 mm.
2527 |
3006 |
< @enesnseccn a | ane _ i 3040 I ~ Ses
3049 |
| 3795
an 67°15 | 180 “ ree , 3829 ,,
3844
1» eesccccccs| 1469 | 183°75 | 1°1636 1628 1894 40°14
"05 1760

800

be uninfinenced.

RAMSAY AND YOUNG: EVAPORATION

At this stage, a small quantity of acetic acid made its appearance
at the lower end of the tube, and passed up, necessarily increasing the
weight. To ascertain the amount of this increase, the readings just
given were repeated, and as the increase was only about 0:2 per cent.,
it was considered that at the same volumes the vapour-density would
The weight was increased from 0°005358 to

0:005368 gram. This weight applies only to this series which was
the last of our experiments. |
Pressure oe te
Jacketing vapour. rte sin em Volume. | Pressure.| P.V. | aude.
* | mometer.
mm. mm
Aniline ..........| 740 °9 183°7° | 1°1636 1630 1897 40°16
ga ERIS « sa 1°0577 | 1764 1866 40°83
” - ” 0 ‘98678 1876 1851 41°15
9 OO Sw Ce cees - rm 0°87902 | 2066 | 1817 41°94
9» 99 - 0°77135 | 2299 | 1773 42 -96
” ” 90 0 66471 2599 1728 44°09
» e » ™ 0 55791 2993 1669 45 °63
» ” * 0°45183 | 3511 1587 48 ‘03
- - mt 0°41677 | 3733 1556 48 °97
9” : - 0-38181 3932 1502 50°76
(liquid)
Pressure of Aniline.
,
Pressure en
reduced ae toe Pressure.
to 0°. airther- |
mometer. |
mm.
4149
745 °9 mm. 183 -7° pon Mean 4152 mm.
4167

AND DISSOCIATION.

Pressure of Methyl Salicylate.

Tempera-
Pressure ture on Vapour-
reduced . Volume. Pressure. ) is —
. air ther- density.
to O°. a
| mometer.
mm. c.c. mm.
| 249 °35 180° ~- 3817 — —_
/ j 4751
: e C is _
330 85 190 {| 4f65 } Mean 4758 mm.
432 °35 200 1°1638 1780 2072 38 -02
: | ” 1°0225 1984 2030 38°80
mm 0-84318 2341 1974 39°91
% se 0 °66484 2852 1896 41°53
me 0 *48711 3679 1792 43 -96
- “ 0°38189 4452 1701 46°32
o - 0°31215 5178 1616 48 “73
; 99 0°27723 5607 1554 50°67
(liquid)
" 0 °25975 5760 1500 52°49
(liquid)
” 9 <a { Pe } Mean 5860 mm.
net oe 7211 -.
557 °3 210 ne { on \ » 7210 ,,
8689
710°1 220 - aca: » 8710 ,
| 8723

Pressure of Bromonaphthalene.

Weight = 0°07374 gram.

Pressure Tempera- |
ture on
reduced . Pressure.
5 air ther-
to 0”.
mometer.
mm. mm.
8855
181°75 220° sg79 Mean 8891 mm.
8940
10617
em. 10701
: 236 °6 230 10700 » 10688 ,,
10734 }
12783 )
. 12801
304°15 240 12811 f » SB «
12850

RAMSAY AND YOUNG: EVAPORATION

Weight = 0:005358 gram.

Pressure Tempera- y
reduced cane om Volume. Pressure. P.V. —
to 0° air ther- density.
: mometer. |
mm. c.c, mm. mm.
304°15 240° | 1°1652 2139 2492 34°27
- - 1°0237 2389 2446 34°93
es i | 0-84417 2851 2407 | 35°49
- 99 | 0 °66563 3512 2338 3-54
ia - 0 -48769 4578 2233 38 °26
. - 0°38234 5612 2145 39°81
- os 0°31251 6604 2063 41°39 i
rs a 0° 24254 8043 1950 43°78 |
es 4 0-20747 9006 1869 45°71
- - 0 17233 10295 1774 48°14
°° 7 0°13714 11930 1637 §2°21
— — — 12575 — —
— _— — 12706 —_ —
—- = = ona } Mean 12813 mm.
Weight = 0°07374 gram.
Pressure —
reduced | — on Volume. Pressure.
too. | ther-
c mometer.
mm. | | c.c. mm.
15077
386°35 | 250° - lores ¢ Mean, 15160.
5172
15222
| 17775
487°35 | 2600 | — ty919 Mean, 17892.
| 17965
608 °75 270 1°1839 19327 Liquid?
a Z | 1°1661 19447)
- . | 1°1307 19766 | Liquid present and increasing ;
8 , 1°0954 20077 +} in amount with decrease of
7 ie | 1-06V1 20370 volume.
; “ 1°0245 20638
” : ae sos | Mean vapour-pressure, 21100.
ce . 0°73741 21071 (Obtained graphically).
0 °38274 21241 4

AND DISSOCIATION.

Weight = 0°07374 gram.

Pressure rape Vv
r2duced ae Volume. | Pressure. | Bo¥s aotaiee
to 0°. air ther- | ensity.
mometer. | |
|
mm. | c.c. mm. mm.
750 °5 279 °85° 1°1665 | 2u669 24112 52°64
“7 - 1:1311 | 21043 23807 53 *22
m ns 1:0958 | 21444 23501 53 90
" ; 1 0604 21762 23088 54°86
- pa 10249 22147 22707 55°78
t. » | 0-98930 22565 22328 56°73
" | 0°95343 22993 21929 57°76
: - | 0-91748 23383 21455 59-04
: wi 0°88125 | 23907 | 21077 60 09
Bi 0°84514 | 24260 | 20504 61-77
" + 0-86280 | 24075 20777 60 “97
, a _ 24745 }
on “ — 24005 | Mean 24616 mm.
rs i — 24447
Weight = 0:005358 gram.
prewure | Temper | | —
reduced | gether | Volume. Pressure.. | P.V. density.
too. | |
| mometer. |
mm. c.c. mm. mm.
759°5 | 280°4° | 1°1665 2400 | 2800 32°91
- - | 1-1811 2475 2800 $2 -92
es | es 1 0958 2556 2801 32-89
‘5 99 | 170604 2639 2798 32-93
. a "| 4 -0249 2731 2799 32°91
- na | 0-9893 2824 | 2794 32°97
m . 0°8451 3284 | 2775 33 +20
is " 0 -6664 4085 2722 33 °85
- a 0 -4882 5383 2629 35°06
ts ne 0 “3828 6661 | — 2550 36°14
‘ i 0°2779 8742 | = 2429 37-93
‘ - 0 +2077 11048 2295 40°15
is a | 01725 12742 | 2199 41 ‘92
‘ | - | 0-°1378 15105 | 2073 44°43
as 01020 18496 | 1887 _ 48°84
“ a 0 -08432 20693 | 1745 52°79
‘ 0-07724 21848 | 1689 54°57
0 06664 23483 «=| ~—s(1561 58°99

23891 Very little liquid.
24556
24741 } Mean 24680 mm.
24744

Reduction and Arrangement of Results.
23. Vapour-pressures.—As we have already published two papers

804 RAMSAY AND YOUNG: EVAPORATION

on this subject (Trans. Roy. Soc., 1884, 465; and Ohem. Soc. Trans.,
1885, 42), and have there discussed the results of other observers, we
need not reproduce them here. ‘The measurements were there given
up to a pressure of 720 mm.; above that pressure the results
are new.

We have first to point out an anomalous condition with acetic acid
for which we have discovered no parallel in the cases of alcohol and
ether. With alcohol and with ether, constant results were obtained,
whatever were the relative volumes of liquid and vapour. But with
acetic acid condensation takes place before the pressure becomes
nearly constant; when a considerable amount of liquid is present,
however, pressure remains nearly constant with decrease of volume.
It might at first be supposed that this behaviour is due to the presence
of air, or some indifferent gas which had originally been dissolved in
the liquid; but the complete absence of air is insured by the method
of filling, moreover, its absence is demonstrated by the fact that the
pressure remains nearly constant until condensation is complete;
whereas when air is present, a bubble is left behind, and is not
dissolved without a considerable rise of pressure. It has been stated
by Wiillner and Grotrian (Pogg. Ann., 11, 545) that the phenomenon
is general; this, however, is the first instance in which we have
observed it; and, as our work on alcohol and on ether shows no sign
of it, we cannot agree with their opinion that it is general. The
question may be asked, then—What is the true vapour-pressure ?
We have taken the mean of four fairly concordant observations when
condensation is well advanced. As might be expected, however, the
agreement between individual observations is not so good as with
alcohol and ether.

24. As with alcohol and ether, a formula of the form recommended
by Biot was calculated so as to smooth the results of individual

observations, und to afford a means of calculating the values of ®.

The results of experiment were first smoothed graphically, and five
pressures were chosen corresponding with equal intervals of tempera-
ture, viz., 0°, 70°, 140°, 210°, and 280°. The corresponding pressures
are 3°3 mm., 136 mm., 1414 mm., 7237 mm., and 24620 mm. The
constants for the formula—
log p = a + bat x cf,
are a = 6°7003112 log a = 1:998815139,
log b = 0°6879733 log B = 1:994508740,
log c = 0°1162143,
b and ¢ are both negative.
We now give a table showing the results obtained experimentally,
and those calculated by the above formula. The results previously

given (loc. cit.) which were obtained graphically are under the
heading of found numbers. The value for 0° is extrapolated from

AND DISSOCIATION.

805

2°7°, where the last observation with the liquid acid was made.

25. Vapour-pressures of Acetic Acid.

Pressure. Pressure.
Tempera Tempera- =
ture ture.
Found. Calculated. Found. Calculated.
mm. mm. mm. mm.
0° 3°5 3°30 150° 1846 1846 °8
10 6°34 6°38 160 2370 2381°6
20 11°80 11°73 170 3031 3035°2
30 19°9 20°61 180 3819 3826 °4
40 34°0 34°77 190 4750 4775°5
50 56 °2 56°56 200 5860 5904 °7
60 88 °3 88 °94 210 7210 7237 °9
70 137°1 136°0 220 8801 8800-1
80 202°1 202 °3 230 10688 10619 ‘0
90 292 °8 293 °7 240 12794 12724°0
100 416°5 417°1 250 15159 15144°0
110 582 °6 580°8 260 17892 17913 °0
120 _ 794°0 270 21060 21063 -O0
130 1040°0 1067 °6 280 24603* 24629 -O
140 1381 ‘0 1414°0 me — —_

* Calculated from observations at 279°85° and at 280°4°.

Rate of Increase of Pressure per Unit Rise of Temperature.

26. The values of these constants, e, were calculated in the follow-

ing manner :—The pressures at one-tenth of a degree above, and at one-
tenth of a degree below any given temperature were calculated, and
The results are given in the annexed

the difference multiplied by 5.

table.
Tempera- dp Tempera- dp Tempera- dp
ture. dt ture. dt’ ture. at
mm. mm. mm.

0° 0° 227 100° 14°20 200° 123 °1

10 0 °404 110 18°75 210 144°3

20 0-686 120 24°20 220 168 °5

30 1°120 130 30°80 230 195 2

40 1°750 140 38 “70 240 225 °3

50 2 °645 150 48°15 250 258°5

60 3 895 160 59 -00 260 294 °3

70 5 °550 170 72 00 270 334 °9

80 7°78 180 86°85 280 378 °5
90 10°60 190 103°5 - a

RAMSAY AND YOUNG: EVAPORATION

Expansion of Liquid Acetic Acid.

27. As all the observations, with few exceptions, have been made
at definite intervals of temperature, the tables already given need not
be reproduced.

But it appears advisable to give here a comparison of the results
obtained by us with those of other observers. This, however, is
difficult. Kopp (Annalen, 64, 216, and 92, 11; and Pogg. Ann.,
72, 1, 223) gives a series of determinations of the expansion of acetic
acid; but it is evident, on reviewing his results, that the sample of
acid used by him is not quite comparable with ours. The boiling
point of our acid at 760 mm., calculated from a direct observation at
765°2 mm., is 118°5°; and from our constants, with Biot’s formula,
118°55°; whereas Kopp’s sample boiled at 117°3° at 760 mm. The
sp. gr. of Kopp’s sample at 0° (probably extrapolated) was 1:0801,
while ours, also extrapolated, is 1:0697. Nevertheless, the rate of
expansion between 30° ‘and 120° is nearly the same in both cases, and
we have therefore taken the mean of three comparisons at 30°, 40°,
and 70°, to reduce Kopp’s values, in which the volume at 0° is taken
as unity, to the volumes of 1 gram. Weappend a comparison of the
volumes obtained by him and by us.

Volume of 1 gram. Volume of 1 gram.

Tempera- |____ _}| Tempera- |
ture.
Kopp. | R. and Y. | Kopp. | R. and Y.

| |

9323 | ‘9348 | 1°0052 | 1-°0052
9422 | ‘9440 | 1°0168 1°0168
‘9521 | 9532 10288 | 1:°0290
‘9622 ‘9623 10415 | 1-°0418
‘9725 | ‘9724 1 ‘0547 1°0545
‘9831 | ‘9828 1 °0685 1°0682
*9940 “9940 — —

At higher temperatures, however, Kopp’s formula,

v=1+4+ at + db? + cé’,

where the logarithms of a, b, and ¢ are 3:02408, 7°26300, and 9-98423,
as might be expected, ceases to represent the results. For instance,
at 200° the volume calculated is 1°2081, and found by us 1°2117; and
at 208°, 1°4189, compared with 1°5172.

28. Perkin has recently made determinations of the sp. gr. of acetic
acid (Trans., 1884, 481). His sample, however, did not show a
constant boiling point. The specific volume found by him at 15° was

} | | ,
———t es — —}—____+- iz — 4 008%
oe ee |
+ + ] T
| | bes | |
2 a ae in a rT . 0092
} |
7 rT
pitt ij ty a ™
| | | |
ECCce Ht Mt
| + 7 000'E
| }
Sot
——{-— a 00z'¢
ee Oe ee
|
| | |_|
Sennen -
| | | |
| Reem
|
— ca a a a oe r 009°.
— ~4——+ oe ae oe
| | | ||
a st | YT | 008¢
} | ——+-—+—+ },4
} } | 1]
| A
= a a 2 oe 8 eT | 000%
—— ++ |
Se ||
wy 9g +s = of
‘SWHZHLOSI Sie!

(our d puw o} 38 T= H)
‘dIOV OILLEAOV JO ANOMVA AALVUALVSNO aNV GaLVUALVS JO SAHILISNAG
T id GSEL 4AIGOPIQ) POG’ ULIY) “ULNY]*

‘NVANIG AG SASSOYUO SHL

—

‘OM sue] sureyy ;S “TArT ‘stg x

SNNOA GNV AVSWNVY AG SNOILVNINYSLSG LNASSYdSY SANTONIO SHL

(1-H) Ansuaq
bb 5b

|
wf

908

00%

009

008

000

006

—— 006't

009't

008 '‘t

000%

0022

PANS SAA

AND DISSOCIATION. 807

0°9468, and at 25°, 0°9570. We find 0°9482 and 09579 at these

temperatures.
A review of the determinations of the sp. gr. of acetic acid is given
by Lossen (Annalen, 214, 97).

Vapour-densities of Acetic Acid.

29. The results actually obtained by experiment are represented by
circles on Plate I. The vapour-pressure corresponding with each
isotherm having been marked off, curves were drawn to pass through
the experimental points, which cut the horizontal lines representing
vapour-pressures at the condensation points. The curve which passes
through these points of intersection gives the densities of the satu-
rated vapour at any pressure.

In Plate II, the variation of vapour-density with alteration of
temperature, is shown, at constant pressure. The diagram was con-
structed by reading off from Plate I, and from a similar diagram
giving results at higher temperatures, the densities of vapour corre-
sponding with temperatures of observation at definite pressures.
These were then plotted to give isobaric curves. The temperatures
corresponding with different vapour-pressures are here represented by
straight lines, and the points of intersection of the isobars with these
lines again represent the densities of saturated vapour, but now
referred to definite temperatures. It will be seen on inspection that
Plate II includes the whole of our observations, whilst Plate I
represents the lower portion in detail. It has already been stated in
Section 23, that condensation ‘begins before the vapour-pressure is
attained; this is seen on the diagram (Plate I) where the experi-
mental results, near the saturation point, do not correspond with the
lines drawn.

30. The table (p. 808) shows the varour-densities at definite
temperatures and pressures.

31. The vapour-densities of acetic acid at pressures below an
atmosphere have been made the subject of experiment by Naumann
(Annalen, 155, 325). Naumann’s measurements were made in a
Hofmann’s ‘ube, but without any arrangement for controlling
pressure; the pressure adjusted itself according to the temperature
and the weight of substance taken. His temperatures, also, have
confessedly no great claim to accuracy, for the jacketing vapours
used at temperatures between 100° and 185° (which were obtained by
water-vapour and aniline-vapour boiling at a pressure which he does
not give), were those of various mixtures of hydrocarbons. His
results, therefore, can only claim to be approximate; as they serve,
however, to corroborate our results, we give a few of his measure-
ments for the sake of comparison (p. 809).

TY 30 Aptsaearag

UvSTTO

See ia SATS
slatctn oe “> he OD ee achetin=
7 04-09 | %&-1S GP. oF 01-68 99-8 1Z- 28 082
S 00-2 | 08-2h | 68-98 | 68-28 09%
a 09.2% 64-88 00-8 OFZ
P| 68: ZF 28-8 022
rs) 66- LP £9.88 002
| 8S. ZF 68-L4E 10. 8 99-28 LI. 18 O8T
> 18. LP 08 - 1F 88-9 LL-¥8 OT- Ze 09T
) 18: 9F ZP- IF GO- 88 ZL. $& ZI-28 Ze. 18 OF
* 1Z-9F 6L: ZF 0S. 98 10. ¥€ GL. 28 OZ
i) 0S: 8F 18: IF 26- LE £9.98 00T
S 80-8 | 98-6 | oF-0F | 08
=) 1z- 1S ZI- LP CS. Sh OL
a 88-09 £0- LP 09
= 19- 8¢ Z9- OS 0g
- Zz- ¥S oOF
a ‘ ‘ ‘
a 00062 000‘0Z 000‘9T 000‘0L 000‘ 0002 000‘T 00s 008 Oot OF 02
*s0in4
< -e10d ute J,
[= =}

ie)
S
-¢)

"BIOJOUIT][IUL UI saInssolg

AND DISSOCIATION. 809

Temperature.
Pressure. | l l
> Naumann. | R. and Y. | Naumann.| R. and Y. | Naumann. | R. and Y.
we | 100° 100° 120° | 120°
7” a
100 45°7 48-75 | 40°5 | 41-81 | 35:1 | 36°
180 50 °0 51°95 43 *9 | 45°00 38°8 | 39°56
300 ~ — 47-2 | 48°50 424 | 42-79
400 — — 49 9 50°64 44°7 44°99
Temperature.
Pressure. |
mm. | Naumann. | R. and Y. | Naumann. | R. and Y.
140° 140° 185° 185°

32°0 32°28
32°8 32°9

|
100 | 32-0 33°72 — ome
180 | 34°4 35°92 | 30°9 31°52
300 | |

32. Horstmann also has made experiments on the vapour-density of

acetic acid. In his first paper (Ber., 2, 299), he suggested that a
difference should exist between the forms of the curves representing
the densities of the saturated vapour in the case of a true dissociating
substance and of acetic acid. He supposed that the density of dis-
sociating substances increases with fall of temperature, but that in
the case of acetic acid it decreases; and in his second paper (Ber., 3,
78) he gives experimental determinations of the density of the
saturated vapour, which, using his calculations, confirm his view,
although they are in direct contradiction to the experimental evidence
adduced by Bineau (Annalen, 60, 157) and by us. Horstmann’s
experiments were made by passing air over acetic acid kept at a
known temperature, and estimating by absorption with caustic potash
the amount of acid carried over. He calculated the density by means

of the formula

| G(P—p) |
Vo.p-0°001293 x 14435’
where G is the weight of acid, P the atmospheric pressure, p the
vapour-pressure of acetic acid, and V, the volume of air reduced to
normal temperature and pressure.

33. But in calculating, he employed Landolt’s pressures, which at
the lowest temperatures are nearly double those obtained by Regnault
and by us. On recalculating his results with true vapour-pressures,

it is found that there is a tendency for the density of the saturated
31

VOL. XLIX.

810 RAMSAY AND YOUNG: EVAPORATION

vapour to rise, instead of to fall, with decrease of temperature ;
though the results are not sufficiently concordant to be plotted in the
form of a curve.

34. We append these results as given by him, and, as recalculated,
converting his specific gravities into vapour-densities compared with
hydrogen. For temperatures below 17°, the vapour-pressures of the
solid acid were employed.

| Pressure. Density.
Temperature. |
| Landolt. R. and Y. Landolt. | R. and Y.
mm. mm. |
12°4° 13°5 6°40 27°29 58°27
12°7 13°7 6°51 28°29 | 60°11
14°7 15°1 7°85 25°70 | 49°92
15°6 15°6 8°60 28°58 | 52°33
17*4 16°8 10-00 30°31 | 51°39
20 °2 19°0 11-80 32°91 | 53°51
21°5 20°4 12°72 32°33 | 52°39
22 °6 212 13°50 | 33 ‘06 52°46
25-0 23 °5 15°35 34°93 54°08
26°5 25-0 16°60 33-49 51°01
27°6 26 °5 17 “55 35°51 54°28
33°3 33 “4 23-0 37°24 54°85
38°5 41°5 31-2 39 -26 52°97
38 °5 41 °5 31-2 40°27 52°97
44°6 | 53°1 42-0 | 39°70 50°98
48 °7 63-0 51°5 43 -02 53°50
51°1 | 69-0 580 60 | «45-62 55°14
59°9 | 97-0 88 °3 | 45-04 50°13
62°9 | 1092 99°0 44°89 50 29
63°1 110-0 100°5 | 46-05 51°14

| “ |
These figures speak for themselves as to the value to be attached

to Landolt’s and Horstmann’s experimental work.
35. Bineau’s determinations are as follows :—

] |
ss ga Pressure. | Density. ~— _— | Pressure. Density.

ure, 1

mm. mm.

12° 2°44 54°9 28 -0° 10 03 54:1
19 2°60 52°8 35 °0 11°19 52°5
22 2°70 51°4 36 °5 11 *32 52°3
12 5°23 56°6 20-0 4°0 54°0
20 5°56 54°4 - 5°6 54°4
24 5°7 53°4 om 8°5 56°0
30 6°03 52°0 - 10°0 57 °2
20 8°55 56°0 30°0 6°0 | 52°0
22 8°64 55°6 ™ 10°7 53°8
20 °5 10°03 | 57°0 _ — —

:
a &
ir | 773
Qa f\8 il
2 UN eae,
‘ ,
> =} \P3 | = 7 8
g LUNE COMTI.
eel PANGS S SA ;
© IHN LC AL
: HH AEE az tT
. ii TS) Seen Sseseh Reeds .
ae 2 Sb iti iii ik
a sie A VY ee AN {LA ft ;
CE ee Ua TAG THOTT
abe, LAP VEEL PL EL | ATi a
Se LW T
ce: 71 OTL LTT ~
5 CAA AY EV [ :
Bs 717 LTA MYM TA .
TT TALLMAN TY
ea /l / LE MN MMI LY PL
| LYMM ITO TT AY ;
8 ATTN,
§ LALLY TT: |
3 AY HA.
& {A AZ / / |_ | | | | |
sa oY VV YW) bV/V Uy V| r | rT
8A = Sy 7 / 225 Gs GSE GEE Se GE Ge ee Gee
ne ZeZZYZ Wh] 7 (rT i
sz EEL L|WKA LIA jt 4 | | | | 4,
2A IZZG L|IwAnAS! ft} tty tt ty
§ eS a | | "6 a A. | | | -
ce | AT | TT | | i oe ee &
: | ae a“ : +++ +1 14
ES a a co a ek ae: Sc oe ch Se Se Soe

‘acnyo.4ad ua qT

AND DISSOCIATION. 811

The series at 20° constitutes an isothermal obtained by Bineau
from his observations. A straight line drawn through these points,
to meet the vapour-pressure line, would give the density of the satu-
rated vapour equal to about 58°4; whilst our results would give, if
extrapolated, about 59°3, a very close agreement. Bineau’s vapour-
pressures also agree well with curs.

36. We next give a table of the densities of the saturated vapour,
read from the curve on Plate II. It should be borne in mind that on
account of the abnormal behaviour of acetic acid while condensing,
these numbers must be regarded as minimum values, although above
40° it is probable that they are nearly correct :—

— Density. ™ = Density. _— Density.
20° 59°3 110° 50°60 200° 51 06

30 57°7 120 50 *32 210 51°59

40 56°2 130 50°16 220 52°26

50 55°0 140 50°08 230 53°13

60 53°77 150 50 06 240 54°24

70 52°76 160 50°12 250 55 *62

80 52°00 170 50°22 260 57°28

90 51°40 180 50 40 270 59°43
100 50°95 190 50 64 280 62°62

37. Latent Heats of Vaporisation.

From these densities of saturated vapour, the heats of vaporisation
were calculated by the formula

t
be

where L = heat of vaporisation; S,= volume of 1 gram of satu-

L=(8,—8)*2.

d
rated vapour; S, = volume of 1 gram of liquid ; f = rate of increase

of pressure in absolute units (grams per square centimetre) per unit
rise of temperature at the given temperature; ¢ = absolute tempe-
rature ; and J = the mechanical equivalent of heat (42500).

The data and results are given in the following table (p. 812).

The heat of vaporisation of acetic acid has been measured at its
boiling point under atmospheric pressure by Favre and Silbermann,
and by Berthelot. The former (Annalen, 64, 186) used acid boiling
at 120°, and found the heat of vaporisation to be 101°9 cals.; and
Berthelot (Hssais de Mechan. Chim., 1, 418) gives the number
120°8 cals. Our value at 120° is 92°71. It appears probable that

812 RAMSAY AND YOUNG: EVAPORATION AND DISSOCIATION.

- -
+ S,. (S,—S.). | ¢. . |
af
}
20° 13080 13079 | 0933 0 -00689 84°05
30 7911 7910 | 1°523 0:00713 85°88
40 4970 4969 | 2°879 0:00736 87 “02
50 3226 3225 3 °596 0 °00760 88°14
60 2164 2163 5296 0 -00783 89°69
70 | 1486 1485 7 546 0 00807 90°43
80 | 1043 1042 | 10°58 0 -00831 91°59
90 | 747-5 746 °5 14°41 0 -00854. 91°88
100 545 °6 544°6 19°31 0 -00878 92 *32
110 | 405-1 404°0 | 25°49 000901 92°79
120 | $05°7 304°6 32°90 0 00925 92°71
130 | 233-9 «| = (232-8 41°88 | 0°00948 92-42
140 181°3 180 °2 52°62 | 0°00972 | 92°16
150 142 °2 141°1 65°47 0-00995 | 91°91
160 112°7 111°6 80°22 001019 | 91°22
170 90 -29 89-14 97°89 0 01042 | 90 “92
180 72°97 71°80 118°1 001066 90°38
190 59°48 58°29 | 140°7 0-01089 | 89°32
200 48°74 47°53 | 167°4 0°01113 | 88°54
210 | 40°19 38°95 196°1 0°01136 86 “82
220 33°32 32 06 229 °1 0°01160 85 20
230 27°72 26 -43 265 ‘3 0 -01184 83 -03
240 | 23-12 21 ‘80 306 °3 0°01207 | 80°59
250 19°32 17°95 351°5 0°01231 | 77°65
260 | 16°17 14°76 400 *1 0:01254 | 74°07
270 13°51 12-05 455 °3 0°01278 | 70°12
280 | 11°21 9°69 514°6 0°01301 | 64°88

the higher numbers found directly were due to want of purity of
the samplesemployed. With alcohol and with ether, there was a close
concordance between our results and the direct observations by
Andrews, Regnault, and Favre and Silbermann. With these sub-
stances, also, the heats of vaporisation do not rise with fall of tem-
perature; but with acetic acid there is a rise from 280° downwards,
reaching a maximum at about 110°, and then a continuous fall.
This is shown graphically in Plate III.

For the sake of comparison, small diagrams showing the relation
of the vapour-density of alcohol to temperature and pressure have
been inserted. In Plate IV, isothermals at comparatively low pres-
sures are given; the trend of the curve below 110° is characteristic
of stable substances. In Plate V, a large portion of the isobaric
diagram is given. At higher temperatures than are shown in the
diagram, the curve becomes more and more horizontal until the
critical temperature and pressure are reached. The fundamental
difference between these diagrams and those exhibiting similar rela-
tions of acetic acid is very striking.

rma

Temperature.

—

Journ. Chem.Soc. October 188 6 Plate Ii.
HEATS OF VAPORIZATION OF ACETIC ACID.

Temperature.

Calories.
60 10 80 90 100 on
2804— ~ -
| ne
| N |
266 |4 > of
240" N 20"
220 220°
.
200" oss
ame : 1
; | :
80 +4 as
|
160° 160°
I o
790 '| > 140
D 20°
|
100" |-+-+ _
| |
e
or tT 80
ri} | | | |
| |
| | | |
° iL i | =
60°} Ht | ! bag
|
: # | ia
d
Y ° 5
20° lorves a
60 70 00 90 100

Harrison *& Sons, Lith, St Martins Lane. W.C

Journ. Chem.Soc. October 1886 Plate IV

DENSITIES OF SATURATED AND UNSATURATED |
VAPOUR OF ALCOHOL.

(H =1 at t° and p mms.)

(Lower portion of diagram only)
ISOTHERMS.

WuUNS
EE apenntenanetngemnanenscnnanaiatie
| ee a Se Tee
12,000 | a - ye
6A
71,000 — yy | Z|
| a Sn Se WA
| 160
9,000 = | = e
C—O — a
s | 150
o 7,000 4. ’
| |
Y 6000
A, -~ er | ee
| |
5,000 *— + Ee LL
f— 130
4,000 \- —
| 420
5000; edie i 1
| se
-
2000 FFL 00 —— -
| ° |
1000 |-Fige@ SF |
a |
0 b73° —— i | | |
23 4 25 26 27 28 29
Density ( H-1)

Harnson & Sons. Lith, St Martins LaneW.C

S
4

(

eanqyoaduea [

Harris:

Journ. Chem.Soc. October 1886

DENSITIES OF SATURATED AND UNSATURATED
VAPOUR OF ALCOHOL.

ISOBARS.

(H =1 at t° and p mms.)

Plate V.

24 & 37 JJ 35 37 39
“N
NK | TWN
—
220 — a
_ ——|
ee —
Se P6825 mms
LL
200 £2 18
46-4
180 ‘ni =
14, t
|
160 | }
|
190 onal ’
|
|
120 4 ——
'
Nd —
3 |
5 100
5 — -+
S 80 —
60 }
a | ||
- ; |
| | |
j EE EE a ee }
| rT TT 4 |
29}——+—__+__1 |
| | | | | |
| | i | L Tein ‘2
| | | | | | | |
ttititt! i i} | t | L |
23 25 27 29 3/ 33 JS 57 39

Harrison & Sons, Lith. S' Martins Lane W.C

Densuty (H= 1).

813

LXXVI.—Trimethyldiethylamidobenzene.

By R. F. Rurray, B.A., M.D., Lecturer on Chemistry in
McGill University, Montreal.

In 1872 A. W. Hofmann (Ber., 5, 720) observed that at high tem-
peratures the halogen salts of secondary, tertiary, or even quaternary
bases of the aromatic series are converted into the salts of their
isomeric primary bases by a transfer of the fatty radicles from the
side-chain to the nucleus. Thus the hydriodide of monomethyl-
aniline at a high temperature becomes converted into the cor-
responding salt of toluidine :—

C.H; C.H,(CHs)
CH; pNHI = H >NHI.
H H

This reaction is now developed technically, and is largely employed
by the Gesellschaft fiir Anilinfabrik in Berlin for the preparation of
the higher homologues of aniline, especially pseudocumidine. On
submitting the latter to his reaction, Hofmann has recently obtained
an interesting substance, pentamethylamidobenzene, the first homo-
logue of aniline in which all the hydrogen-atoms of the benzene
nucleus are replaced by fatty radicles (Ber., 18,1821). It is known
that the transfer of radicles from the side-chain to the benzene-ring
is by no means confined to the methyl series. Hofmann himself has
shown that the phenyl-groups in aniline may be ethylated or amylated
in the same manner. It seemed of some interest, however, to ascer-
tain how far the presence of methyl-groups in the phenyl-group of
aniline would promote or retard the introduction of other radicles of
different composition. In order to decide this question, I began by
submitting pseudocumidine to the action of ethyl alcohol, with the
view of obtaining a primary monamine in which the phenyl-group
should contain three methyl- and two ethyl-groups, that is of the
formula C,(CH;);(C:H;)2NH,. The first step in the production of
this compound is the preparation of ethylcumidine. For this pur-
pose dry cumidine hydrochloride was mixed with somewhat more
than the theoretical quantity of ethyl alcohol, and the mixture
heated in a closed tube at 120—130° for four hours. The product
was found to be the hydrochloride of the ethylated cumidine, together
with some ether formed during the reaction. Excess of alkali added
to this product caused the ethylated cumidine and the ether to rise
to the surface, and they were then easily removed and separated.
VOL. XLIX. 3K

814 RUTTAN : TRIMETHYLDIETHYLAMIDOBENZENE.

Distillation very readily removes the ether and leaves a basic residue,
which comes over between 220° and 230°.

It is worthy of note that this method of ethylating cumidine yields
better results than the usual one of digesting the base with a haloid
ether, and then separating the secondary amine by excess of alkali.
Moreover, the employment of the more costly iodide is avoided.

The mixture of secondary and tertiary bases thus obtained is sealed
up in tubes with about an equal weight of ethyl iodide, and heated to
a temperature of 260° to 280° for eight or ten hours. If the operation
be successful, the tubes, on cooling, contain a dark-red jelly-like mass,
which cannot be poured out. This consists of the new base mixed
with other bases, a small quantity of some liquid, aromatic hydro-
carbons, and a tarry substance. The hydrocarbons are removed by
distillation with steam, and the residue filtered and treated with con-
centrated alkali, when a thick reddish strongly basic oil rises to the
surface. This is fractionally distilled, and the part which comes
over between 285° and 290° is retained and treated with hydro-
chloric acid, when a very insoluble hydrochloride, crystallising in
groups of needles, is at once formed. These are not soluble in cold
water to any appreciable extent, and only very slightly soluble in
boiling water, but easily soluble in alcohol.

After one or two recrystallisations, this salt was obtained in a state
of purity. Analysis proved it to be the hydrochloride of trimethyl-
diethylamidobenzene, C,(CH;);(C,H;)2.NH2,HCl, which requires the
following values :—

Experiment.
Theory. L II. ll.
Oye cccccecs 156°0 67°20 66°82 —_— —
HH» -s ee eeee 22°0 9°65 9°72 — —
). Terre ey 140 7°47 — — —_
Ol ccccvces 35°5 15°68 _— 15°6 15°573
227°5

The free trimethyldiethylamidobenzene obtained from this salt by
the action of an alkali is a liquid boiling between 288—290°. It has a
sp. gr. of 0°971, is quite colourless when first set free, and has a faint
odour. After a time, it becomes thick and dark coloured.

The hydrochloride does not yield a well crystallised platino-
chloride, but with palladium chloride it gives a beautiful green palla-
dium salt in feathery crystals. The acetate and sulphate crystallise
in needles and are very soluble. The oxalate is very sparingly soluble,
and consists of large prisms. The acetyl compound, easily obtained
by the action of acetic anhydride, crystallises in rosettes of needles
and melts at 182°.

STERN : ACTION OF BROMINE ON PHOSPHORUS TRICHLORIDE. 815

The primary character of the base is at once proved by its giving
Hofmann’s isonitrile reaction with alkali and chloroform. As proved
by the researches of Froelich (Ber., 17, 1801 and 2073), the methyl-
groups in pseudocumidine hold the positions 2, 4, and 5, that of the
amido-group being 1. As there are thus only two places vacant for
the ethyl-groups, the substance I have described is represented as
follows :—

NH,

C.H; ~ , CH;

CH, y O.H;
CH,

Chemical Laboratory,
University of Berlin.

LXXVII.—On the Action of Bromine on Phosphorus Trichloride.

By A. L. Srern, Scholar in the Chemical Laboratory of the Mason
College, Birmingham.

Tue action of bromine on trichloride of phosphorus was first investi-
gated by Gladstone (Phil. Mag., 35, 1849, 353—354), when attempt-
ing to form a compound of phosphorus containing both halogens, that
is, a chlorobromide of phosphorus. He found that when bromine was
poured into phosphorus trichloride, it sank to the bottom and two
layers were formed, the upper consisting of a solution of bromine
in the trichloride and the lower of a solution of the trichloride in
bromine. These two layers could not be made to mix, but on adding
a little iodine combination immediately took place with develop-
ment of much heat, and on cooling red crystalline masses separated,
resembling phosphorus pentabromide with excess of bromine. These
crystals having been freed as far as possible from the mother-liquor,
were analysed by decomposing them with water and precipitating
with silver nitrate. Their percentage composition was found to
be—

BREED co ccccccsvcscccccesece 88°6
CRIGTIMO cc ccccccccccccsccccece 6°4
Phosphorus (by difference) ....... 50

This Gladstone regarded not as a compound of both halogens, but

816 STERN ON THE ACTION OF

merely as phosphorus pentabromide contaminated with a little
chloride, the pentabromide requiring—-

BREOMEIRG 6600 cc cccccece 92°45 per cent.

eee eeereeeee ”

The reaction is probably
PCI, + Br,I = PBr,; + IC\,.

Gladstone also tried other methods for forming the chloro-bromide
but without success.

Wichelhaus (Annalen, Supp. 6, 277) states that phosphorus tri-
chloride and bromine combine with development of heat, and that the
product when cooled by a powerful freezing mixture deposits crystals,
which he assumed to be PC],;Br,. On exposure to the ordinary tem-
perature, however, these melt, and the liquid separates into two
layers.

Friedel and Ladenburg (Bull. Soc. Chim., 8, 1867, 146) used a
mixture of PCl,+Br., and found it reacted as if it was PCI,Bro.

Michaelis (Ber., 5, 9) found that the chlorobromide was stable at
ordinary temperatures. By exposing a mixture of the trichloride
with bromine in the proportion PC); : Br, in a tube for some time
to a winter temperature, he obtained the compound PCI;Br, (the
analysis gave PCl,Br, + 5°2 per cent. PCl;), decomposing into PCI,
and Br, at 35°. On placing crystals of this or even of some of the
mother-liquor in a mixture of phosphorus trichloride and bromine,
the lower layer immediately solidified, the compound PCI,;Br, sepa-
rating out, just as a supersaturated solution solidifies when a crystal
of the salt is dropped into it. On treating this compound with PC],
he found it to be unaltered.

Prinvault (Compt. rend., 74, 868) found that a homogeneous liquid
was obtained on gradually adding bromine to phosphorus trichloride.
On heating this on a water-bath until bromine was no longer evolved,
a red liquid was left, which crystallised at —4° or — 5°, forming brown
oily needles, and it is stated that this compound can be distilled un-
changed below 90°, producing a colourless vapour of the formula
PCI;Br;. Quick distillation above 90°, however, decomposes it, crys-
tals of PCl,Br, being obtained in the receiver; if this is dissolved in
cold PCl;, and the solution boiled, yellow crystals of PCl,Br separate
on cooling.

Prinvault also obtained PC],Br, thus :—

PCI,Br, + PCl, = PCl,.Br; + PCI],Br,
PCI,Br;, + PClBr = 2PC1,Br;.

Michaelis (Ber., 5, 414) has repeated the above experiments, and

BROMINE ON PHOSPHORUS TRICHLORIDE. 817

obtained different results. When 2Br, was added to PCI, there
appeared after strong heating a homogeneous oily liquid, which on
cooling crystallised, giving brown needles with a green reflex, of
the composition approximately represented by the formula PC],Br, ;
he found that this compound could not be distilled without decompo-
sition, and that its vapour was of the colour of bromine. Michaelis
considers these compounds to have a similar constitution to PC],,ICl
and PC\;,Fe.Cl,; while Prinvault holds them to be compounds of PBrs
with BrCl.

With a view to the more complete investigation of this reaction,
the following experiments were performed.

Phosphorus trichloride was mixed with bromine (free from iodine)
in various proportions, the mixture allowed to stand, and the crystals
which were deposited were analysed.

Three series of experiments were performed.

Phosphorus trichloride and bromine were run from two burettes into
small stoppered bottles; heat was developed, and on cooling crystals
were deposited ; in order to obtain them free from the mother-liquor,
the contents of the bottles were poured into funnels, the stems of
which were plugged with asbestos, and the funnels were then placed
in a closed vessel over soda-lime. In order to obtain them even
approximately free from the mother-liquor, they had to drain for at
least 24 hours; during that time decomposition seemed to take place,
and on attempting to transfer them from the funnel to asmall weighing
tube, they deliquesced so rapidly that it was useless to analyse them.

A process had therefore to be devised by means of which the
crystals could be freed from the mother-liquor and weighed without
exposing them to the atmosphere. This was done as follows.

A glass tube about } inch in diameter and 10 inches long, was
drawn out into a narrow neck at one end and the bottom of this sealed
off. The narrow neck having been plugged with asbestos, measured
quantities of phosphorus trichloride and bromine were run in at the
top from a dry burette, the lower sealed part of the tube being
occasionally cooled to moderate the heat produced by the reaction,
otherwise the PC], boiled.

The tube was then sealed off at the top. In order to separate the
crystals which formed on cooling, the two ends of the tube were
broken and a current of air dried by sulphuric acid was drawn through
the tube by means of a filter-pump in communication with the lower
end. When the mother-liquor had been drawn off, both ends of the
tube were re-sealed, and the whole weighed.

The crystals were then dissolved in solution of sulphurous acid by
breaking off one end of the tube under the surface of the solution,
and allowing the liquid to rise in the tube. The tube was then

3K 2

$18 STERN ON THE ACTION OF

washed, dried, and re-weighed, together with the ends broken off.
The difference in weight gave the amount taken for analysis.

The solation was made up to a definite volume, and to a measured
quantity of this, decinormal silver nitrate added as long as a precipi-
tate formed. _ The phosphorus was determined by difference.

The trichloride and the bromine were mixed in four different
proportions.

(i.) PCl, : Br,.—Heat was developed on mixing, and two layers
were formed, the lower dark-red and the upper light-red; after
standing a few hours the lower one suddenly solidified to a brick-red
crystalline mass.

(ii.) PCl,; : 2Br,.—More heat was developed on mixing than in the
former case, and a homogeneous red liquid was formed; a quantity
of dark-red crystals separating out on cooling.

(iii.) PCl,; : 3Br,—A considerable amount of heat was developed
on mixing, and a homogeneous red liquid was formed; on sealing up
and shaking, a quantity of dark-red crystals separated.

In each of these three experiments, the mother-liquor was drawn
off on the following day ; temperature about 13°.

(iv.) PC]; : 6Br,—A considerable amount of heat was developed
on mixing; a homogeneous red liquid was formed, and dark-red
crystals separated out on cooling. These were left in contact with
the mother-liquor for 31 days, in order to see if all the chlorine could
be displaced by prolonged contact with excess of bromine. This,
however, was not the case.

As the phosphorus in this series was merely estimated by difference,
and as no precautions were taken to eliminate moisture from the
bromine, or phosphorous acid from the trichloride, and as the use of
asbestos as a filtering material might be open to objection, a second
series of analyses was performed. In this series, the bromine was
first tested for iodine, and found to be free from it, then shaken up,
and left in contact with strong sulphuric acid, and finally distilled
into a dry tube similar in construction to the one previously described,
except that the asbestos was replaced by platinum gauze fused into
the tube. The phosphorus trichloride was also distilled directly into
the tube.

In the former series, the liquids were measured from a burette, in
this one by marking lengths on the tubes proportional to the required
volume.

The mode of analysis was slightly modified; the crystals were
decomposed as before by dilute sulphurous acid, and the solution
made up to a definite volume. To determine the chlorine and bromine,
silver nitrate was added to a portion of the solution and the precipitate
of Ag(Cl,Br) weighed; a portion of this was then transferred to a

BROMINE ON PHOSPHORUS TRICHLORIDE, 819

weighed bulb-tube and heated in a stream of chlorine until the weight
was constant. From these results, the quantity of chlorine and
bromine could be calculated. To another portion of the solution,
bromine in excess was added to oxidise the phosphorous acid to
phosphoric acid, and the phosphorus estimated from the weight of
magnesium pyrophosphate obtained on precipitation with magnesia
mixture.

The bromine and trichloride were mixed in approximately the same
proportions as in Series I, with the exception of the first and last
experiments.

(i.) PCl;: Br.—Two layers were formed; but only a small quantity
of crystals separated at 4°; these were insufficient in quantity to
analyse.

(ii.) PCl;: Br,—Two layers were formed, and a small quantity of
crystals separated at 4°. The mother-liquor was drawn off at this
temperature, after the mixture had stood 16 days at the summer
temperature. The crystals were light red.

(iii.) PCl,:2Br,.—A homogeneous mixture was formed, and a crop
of small light-red crystals separated out at the ordinary summer
temperature; the mixture was cooled to 4° when a second crop of
light-red needle-shaped crystals separated. The mother-liquor was
then drawn off.

A second quantity of bromine and PC], was mixed in approximately
the same proportion. After this had stood for one day at the summer
temperature, the mother-liquor was drawn off; the crystals were
light red, and decomposed on keeping.

(iv.) PCl;:3Br,—The mixture solidified at 4°; on standing for
three days at the summer temperature some crystals were deposited,
the mother-liquor was then drawn off.

(v.) PCl,:4Br,.—After standing a few hours this was cooled to 4°,
when it solidified; the temperature was then gradually raised to 15°
and the liquid portion drawn off: the substance thus obtained solidified
on cooling with ice to a brick-red mass, commenced to darken at 10°,
and fused at 13—17° to a dark-red liquid.

The following conclusions may now be considered as established :—

(i.) When bromine is added to phosphorus trichloride, two atoms
of bromine may be made to unite with one molecule of the chloride,
forming the chlorobromide, PC],;Br, (Michaelis) ; when more bromine
is added one atom of chlorine in this is displaced by bromine, forming
PCl,Br; (see Table II, Series II, No. 1, and Series I, No. 2). Still
more bromine being added, this compound unites with part of it,
forming PCl,Br,; (Table II), and at a lower temperature with still
more bromine.

(ii.) The number of atoms of halogen with which one atom of

STERN ON THE ACTION OF

=)
n
a)

94- OOT - ¢¢-O0T TZ: TOT ¢9- OOT | €Z- 10T _
LT-64 ied vE-8 FE. €8 OT- 8Z ST- 69 1g
L6-¥T —- 20: 61 64-11 8L-FT 9Z- &Z 10 II setog ‘spuyshtc
29-9 = 61-9 80-9 LLL 28-8 d
‘dui94 yey} 48 ‘dure, yey}
Jo UMtIp uoYy sum qv yo uUMvIp 0Y4 ".p 78 FO umvrp
aonby-19eyqjour = 83 — -- sem sonbry-tay3out vay} sea tonbry-coqyjou ayy “dute4
‘dmoz19uuns ye Lep oy} “due; aouutns Jourutns 48 sfep QT pooys oanyxipT
T pooys *agqz + "10g ye sX{ep vary} poo
oe 29-58 ~~ PV. 8L 22 -FL L¥-%9 ig |) ‘*** (eouosegtp
-— €8- 6 -” T€-€T 99-ST 9-42 10 fq snaoydsoyd)
= g¢.¢ = 3-8 | ZI- OT 68-01 d I semeg ‘spuyship
a
‘sep Tg “ep ouo | ‘ep ouo ‘ep ouo .
pooys o1Nj XI] pooys OIN4XIPY = | «Ss POOF OANAXTIPY =| Ss: COONS OANGXITL O SI—8t “dwoay,
|
as 8r-48 | T&-28 | €L- LL 6-69 8L- 8S ig oo 00 00 s+° gpmme
— OL-6 OL-&T ce. LT 8Z -€Z 08- SE 10 -xIuI JO UOoTjIS
_ Z8-% 66: € | Z0-¢ 84-9 ZP-OL Z -odmoo poyepnoyeg
—
“G1Od | Sig *10d | ag*10d g*10d “ag *10d

‘I alavy,

BROMINE ON PHOSPHORUS TRICHLORIDE. 821

Taste II.—Atomic Proportions.

P 1 1 1 1 1
Mixture .... Cl 3 3 3 3 3
Br 2 4 6 8 12
P 1 1 1 om 1
ae. J cl | 2-03 1°35 1°41 an 1°55
“|| Br | 2-29 2-84, 3°68 am 591
P 1 1 1 1 an
——.. / Cl 2-30 1°66 1-69 2°50 -
|! Bre | 3-04 3°89 5°31 7°80 ae
P| 41 1 1 1 1
Mixture coved Cl+Br) 5 7 9 ll 15
Crystals, { - i & 1 1 _ 1
Series I... Cl + Br | 4°32 4°19 5°09 — 7°46
Crystals, r.. © 1 1 1 ~-
Series IT .. Cl + Br | 5°34 5°55 7°00 10°30 —

phosphorus can combine depends on the temperature: thus phosphorus
in presence of 15 atoms of halogen at temperature 13—15° can
combine with only 7 atoms, whereas P in the presence of 11 atoms of
halogen united with 10 atoms to form a compound dissociating
above 10°.

(iii.) All the chlorine in the trichloride cannot be displaced by
the unaided action of bromine, even if such a large excess as 12 atoms
of bromine to 1 mol. of PCl; be allowed to react for a month (Series
I, No. 4); the presence of a small quantity of iodine, however, will
enable a much larger quantity of the chlorine to be displaced
(Gladstone).

Michaelis regards the chlorobromides in question as molecular com-
pounds of PCI, with ClBr, whilst Prinvault considers them to be
molecular compounds of PBr; with ClBr. Michaelis’s view is, how-
ever, now scarcely tenable, as it has been shown above that the Cl in
PCI; may be partly displaced by bromine; Prinvault’s view also does
not seem tenable, as compounds have been obtained with less than
5 atoms of bromine.

It has been shown that the compounds containing the largest amount
of halogen are decomposed at ordinary temperatures, whilst those con-
taining less halogen are stable at this temperature, the compound
PCI],Br, being indeed stable below 35° (Michaelis); and as it is also
known that phosphorus pentachloride is decomposed on vaporisation
into PCl, and Cl, it would seem that the valency of phosphorus in

822 BRIERLEY: THE ELECTROLYTIC PREPARATION

these compounds is a function of the temperature, increasing as the
temperature falls; and that these compounds are simply compounds
of phosphorus with varying quantities of halogen.

In conclusion, I must express my thanks to Dr. Tilden for his kind
and valuable help.

LXXVIII.—The Electrolytic Preparation of Vanadious Sulphate.
By J. T. Briertey, Dalton Chemical Scholar, Owens College.

ProressoR Roscog observed that when a solution of vanadium tri-
oxide was prepared by reducing the higher oxides dissolved in dilute
sulphuric acid, the resulting liquid had sometimes a brown and some-
times a green colour. The brown-coloured liquid is usually obtained
when a solution of vanadium dioxide from which the excess of acid
has been removed by zine, is allowed to absorb oxygen from the air,
being thereby converted into V.0;. That these two compounds are
of the same degree of oxidation has also been shown by Roscoe.

These experiments were repeated very carefully, and it was found
that the green trioxide solution could be at once converted into the
brown by the addition of zinc oxide; this of course removing the
excess of free acid. The resulting brown liquid had still an acid
reaction, but when a large excess of zinc oxide was added it became
colourless, and ultimately the whole of the vanadium was completely
precipitated from solution.

This behaviour of a solution of the trioxide with zinc oxide indi-
cates that the green solution is an acid salt, whilst the brown is
probably basic, a conclusion supported, moreover, by the fact that a
few drops of acid turned the brown liquid green again.

I endeavoured to isolate a pure solution of the trioxide by reducing
a pure solution of V,0; in dilute sulphuric acid with metallic zinc,
until the deep green colour of the trioxide was produced. This
green solution, filtered and mixed with alcohol, deposited an intensely
green oily liquid which could be separated from the lighter watery
portion by means of a funnel. I found, however, that it was im-
possible to separate the vanadium salt from the zinc sulphate by long
continued washing with alcohol, or by any other means tried. I then
endeavoured to obtain by electrolysis a solution of V,0;in sulphuric
acid uncontaminated with any foreign substance. This method was

OF VANADIOUS SULPHATE. 823

found to answer satisfactorily, the apparatus being arranged so that
all the hydrogen was utilised and the oxygen carried away without
interfering with the reduction. Fig. 1 shows the arrangement
adopted. The reduction is best commenced on a strong pure
solution of V.O, in dilute H,SO,, which is readily obtained by the
reducing action of SO,. The deep blue tetroxide solution is placed in
the platinum dish (b) forming the negative electrode of the battery
(d), whilst the porous cell (a) containing the positive electrode of pla-

Fre. 1.

tinum (c) and dilute sulphuric acid just touches the surface of the
liquid. The reduction proceeds regularly and may be accelerated by
heating. In this manner, the reduction and evaporation can go on
simultaneously. The end of the reaction is determined by the intense
green colour of the solution, a drop of which on being taken out and
shaken up with a small quantity of water gives a green liquid with
no tinge of blue. ;

The concentrated green trioxide solution is then mixed with about
twice its bulk of strong sulphuric acid, and allowed to remain for
24 hours; at the end of that time the whole of the trioxide separates
in combination with sulphuric acid as a pale green sandy crystalline
precipitate, which must be placed upon a porous tile, washed with
strong alcohol until the free sulphuric acid is entirely removed, and
then dried over sulphuric acid in a desiccator filled with coal-gas.
The following analyses made with two preparations show that this
sulphate of vanadium trioxide has the formula V,0;(SO;), + 9H,O :—

824 ELECTROLYTIC PREPARATION OF VANADIOUS SULPHATE.

Calculated. Found. Found.

0 23°78 23°93 23°81
Babssseveeds 50°60 50°89 50°82
MD évesiees 25°62 25°18 25°37
100°00 100°00 100°00

The vanadium was estimated as V,O; by heating the salt in a
crucible to bright redness, a fine crystalline residue of V.O; remaining.
The degree of oxidation was confirmed by means of a standard solution
of permanganate, 1 c.c. of which was equal to 0°000677 gram of
oxygen; 0°1120 gram of salt required 7°9 c.c. = 0°0053483 oxygen.
The oxygen required by theory being 0°00566. The variation which
is equal to 1 per cent. is only very small when the readiness with
which the salt absorbs oxygen is considered.

Sulphate of vanadium trioxide dissolves in water, yielding a bright
green solution, from which alkalis precipitate the hydrated trioxide as
a dirty green somewhat gelatinous mass, which is rapidly oxidised in
the air.

The fact that the brown trioxide solution is more basic than the
above was ascertained by the fact that this green hydrated trioxide
readily dissolves in a solution of the foregoing green sulphate, forming
a deep chocolate-coloured liquid, and on evaporation this liquid yields
a shining green amorphous mass, which again dissolves in water,
forming a brown solution.

This method of reduction may also be employed for the prepara-
tion of the other lower oxides of vanadium. Thus a lavender-
coloured solution of vanadium dioxide is obtained on continued
reduction, and from this the hydrated dioxide may be precipitated by
ammonia as a deep purple gelatinous mass.

It is, however, impossible to isolate these compounds in the solid
state on account of their extreme instability.

\B)

LXXIX.—On Ammonia-derivatives of Benzoin.
By Francis R. Japp, F.R.S., and W. H. Winson, Ph.D.

‘Tue action of alcoholic ammonia on benzoin was studied by Laurent
and later by J. Erdmann (Annalen, 135, 181). Both these chemists
obtained a substance of the formula C.,H,,N,0, to which Laurent
gave the name benzoinam. Erdmann further isolated a substance
which he called benzoinimide, assigning to it the formula C,,H,,N.
Lophine was also formed.

The results which one of us had obtained in revising the work of
Laurent and others on the ammonia-derivatives of ‘benzil, appeared
‘to render it desirable to re-investigate the action of ammonia on
benzoin.

The following is the method employed by Erdmann in preparing
the above compounds. Benzoin was heated with an excess of
alcoholic ammenia in sealed tubes at 100°, for four to six hours.
During the heating, the liquid deposited silky needles of benzoinam,
which were afterwards separated by filtration. On allowing the
mother-liquor to evaporate spontaneously, it yielded a yellow crystal-
line mass, which by treatment with boiling alcohol could be separated
into two portions, an insoluble yellow crystalline powder (Erdmann’s
benzoinimide, which was analysed in this condition, without further
purification), and a soluble portion, containing lophine tegether with:a
granular substance, the latter being obtained in quantity insnfficient
for investigation.

We conducted the operation in the above manner and observed the
separation of benzoinam as described by Erdmann. According to
Erdmann this compound is insoluble in aleohol: he therefore con-
tented himself with boiling the product with alcohol and analysing
the supposed insoluble residue. We find, however, that by boiling
with a large volume of alcohol the substance dissolves; the ‘solution
may then be concentrated by distilling off the alcohol, without any
separation of substance, and on standing, the concentrated solution
deposits the benzoinam in tufts of long silky colourless needles. This
process of purification was repeated antil a sample of the product,
when treated with concentrated sulphuric acid, no longer gave a red
coloration, thus denoting the absence of Erdmann’s benzoinimide.
The melting point could not be determined, as the substance decom-
posed on heating, and melted at temperatures varying from 190° to

220° according as the temperature was raised slowly or rapidly.

VOL. XLIX, 3 L

826 JAPP AND WILSON ON

Benzoinam dissolves much more readily in boiling benzene than in
alcohol. Laurent’s statement that it is soluble in water is incorrect.

Analysis of a specimen dried at 120° confirmed the formula
C.,H.N.O given by Laurent and by Erdmann. We quote the
figures, as Erdmann’s analyses were made with benzoinam which had
not been purified by recrystallisation.

Substance. CO.. H,0.
B eteseene 0°1882 0°5726 0°1026
Be benecuns 0°1988 0°6040 0°1073

IIT. 0°1273 gram, burnt with copper oxide in a vacuum, gave
13°79 c.c. of a mixture of nitrogen and nitric oxide, measured dry at
12°8° and under 409°7 mm. pressure. After absorption of the nitric
oxide, there remained 13°79 c.c. of dry nitrogen at 128° and under
3962 mm. pressure.

Calculated for Found.
Cy,H2,N,0. r A =
rn } IL. III.
> eer 336 83°17 82:98 82°86 —
a cueies 24 5°94, 6°06 6:00 —
Shaan weed 28 6°93 — — 6°89
Se 16 3°96 -- —- —
404 100°00

The yellow mother-liquor from the tubes from which the ben-
zoinam had been separated was evaporated to a small bulk, allowed
to crystallise, and the yellow crystalline substance removed by filtra-
tion. When, however, following Erdmann’s directions, we attempted
to separate this product, by boiling with alcohol, into a soluble and
an insoluble portion, we found that by employing sufficient alcohol
and boiling sufficiently long the whole dissolved. The solution thus
obtained, when concentrated and allowed to stand, deposited two sorts
of crystals—yellowish needles of Erdmann’s benzoinimide, which gave
the characteristic blood-red colour with concentrated sulphuric acid,
and granular crystals (Erdmann’s granular substance), the latter
separating much more slowly and adhering for the most part to the
sides of the flask. A mechanical separation was effected by shaking
the contents of the flask before pouring on to the filter, the needles
remaining suspended in the liquid whilst the granular crystals at once
sank. The latter were put on one side for subsequent examination.

The so-called benzoinimide (we shall show later on that this term is
®& misnomer) is not, therefore, as Erdmann asserts, insoluble in alcohol,
but only sparingly soluble; itis in fact considerably more soluble than

AMMONIA-DERIVATIVES OF BENZOIN. 827

benzoinam and may be readily recrystallised. Boiling amyl alcohol
dissolves it still more easily, and deposits it on cooling. The best
solvent, however, is boiling benzene, which dissolves it very readily,
and deposits it in lustrous, flat, obliquely truncated needles.

Erdmann obtained the substance, as already mentioned, in the form
of a yellow crystalline powder, and we found that this yellow colour
could not be removed merely by recrystallisation from:alcohol. The
substance is, however, colourless when pure. The method which we
employed in purifying it is again in contradiction ‘to Erdmann’s state-
ments as to the solubility of the substance. According to Erdmann,
it is insoluble in an alcoholic solution of hydrochloric acid. It is true
that a dilute solution of gaseous hydrochloric acid in alcohol does not
dissolve the substance appreciably more than pure alcohol, but
saturated alcoholic hydrochloric acid dissolves it very readily in the
cold,* and the substance may be precipitated from the solution in
colourless needles by the addition of alcohol. One advantage of this
method is that as benzoinam, lophine, and benzoinidam (vide infra)—
the other substances formed in the benzoin-ammonia reactions—are
readily soluble in dilute alcoholic hydrochloric acid, they remain in
the mother-liquor. The benzoinimide thus purified was finally
recrystallised from a large volume ‘of boiling alcohol,f and was thus
obtained in slender, very lustrous, flat needles, which melted at 246°
(no melting point is given by Erdmann), sublimed at a higher tem-
perature without decomposition, and gave the characteristic red
coloration with concentrated sulphuric acid.

The analysis of the compound dried at 120° gave results agreeing
better with the formula C,,H,N, than with the formula C,,H,,N,
assigned to it by Erdmann :—

Substance. CO. H,0.
B sevccnvune 0°1076 0°3439 0°0518
BE ésnceveees 02328 0°7450 0°1123
EEE sevcevesees 0°2323 0°7429 0°1115

IV. 0°2220 gram, burnt with copper oxide in a vacuum, gave
22°63 c.c. of a mixture of nitrogen and nitric oxide, measured dry at
12°5° and under 464°6 mm. pressure. After absorption of the nitric
oxide, there remained 22°63 c.c. of dry nitrogen at 12°5° and under
443°] mm. pressure.

* The solution is yellow, and if very concentrated speedily deposits a yellow
crystalline substance—probably an unstable hydrochloride—which by exposure to
the air, or by contact with alcohol, becomes white, and is converted into the
original compound.

+ At the time this specimen was prepared, we were not aware that the substance
was soluble in benzene.

342

JAPP AND WILSON ON

€aleulated for Calculated for Found.
CosHopNs C,,Hy,N “ A is
(Japp & Wilson). (Erdmann). I. II. IIE. IV.
C.... * 87°50 87°05 ‘87:17 87°28 87°22 —
wae 5°21 5°70 5°35 5°36 5°33 —
i eed 7°29 7°25 — —- —- 7°31
(100-00 100°00

The percentage of carbon found is too high for Erdmann’s formula,
and that of hydrogen too low; and this applies also to Erdmann’s own
analysis, which: gave C 87°31, and H 5°61 per cent. Erdmann was
doubtless led to: adopt the simpler formula by the ease with which
it enabled him to explain the formation of the substance as a con-
den: ation of ammonia and benzoin with elimination of water, whereas
the formula to which we give the preference requires the simultaneous
abstraction of hydrogen :—

2C,,H»O, + 2NH; = CxHa»N, + 4H,0 + H,.
Benzoin.

The reaction corresponds with that in which diphenanthrylene-
azotide is formed from phenanthraquinone and ammonia, except that
in the latter case, as phenanthraquinone is an analogue of benzil and
not of benzoin, abstraction of oxygen occurs instead of abstraction of
hydrogen :—

2C,,H,O, + 2NH; = C.,H,N. + 3H,0 + O.

Phenanthraquinone. Diphenanthrylene-
azotide.

Erdmann’s compound has the same relation to benzil that diphen-
anthrylene-azotide has to phenanthraquinone. :—

C,4H»O2 (Benzil). C,,H,O, (Phenanthraquinone).
C.,H»)N, (Erdmann’s compound). .C,sH,,N, (Diphenanthrylene-
azotide).

Following out this analogy, we propose to name Erdmann’s com-
pound ditolane-azotide. The name given to it by Erdmann—benzoin-
imide—is misleading: the compound cannot be an imide.

In ditolane-azotide, two tolane-groups are united by two nitrogen-
atoms; in diphenanthrylene-azotide two phenanthrylene-groups are
united by two nitrogen-atoms. The very great stability of these
compounds renders it probable that the nitrogen-atoms form part
of a closed-chain complex. Their constitution might therefore be
expressed by the following formule—

C,H; C—N—C-C,H; C,.H,-C—N—C-C,H,
ost N—O-OH, O.H-0 NOt,

Ditolane-azotide. Diphena:.thrylene-azotide.

AMMONIA-DERIVATIVES OF BENZOIN. 829

in which the mode of distribution of the available affinities in the
closed-chain complex is left undecided.

There are various points of resemblance in the character and
behaviour of these two compounds. Both are substances of very
sparing solubility (diphenanthrylene-azotide is practically insoluble
in the ordinary organic solvents of low boiling point): Both have a
high melting point. Both sublime at a high temperature in flat
lustrous needles, and both give characteristic colour reactions with
concentrated sulphuric acid—ditolane-azotide giving a splendid red
colour, diphenanthrylene-azotide a deep blue. The-more sparing
solubility, the higher melting point, and the higher temperature of
volatilisation, are, as usual, on the side of the diphenylene-derivative
as compared with the dipheny!-derivative.

The fact that ditolane-azotide is the most stable of the ammonia-
derivatives of benzoin rendered it probable that the yield of this
substance might be increased by allowing ammonia to act upon
benzoin at a higher temperature. We found that this result might
be very readily and effectually attained, and the substance prepared
in any desired quantity, by heating benzoin with.ammonium acetate.
100 grams of glacial acetic acid were heated ina flask over a free
flame, adding solid ammonium carbonate until the acid was neutralised
and ammoniacal vapours began to be given off. 1V0 grams of ben-
zoin were then added, and the heating was continued until the whole
of the ammonium acetate had been driven off and the benzoin-
derivative began to volatilise. The flask must be shaken the whole
time to prevent cracking. The melted product, without allowing it
to cool, was poured in a thin stream. into excess of alcohol, ground
in a mortar with the alcohol, and after complete disintegration,
extracted with boiling alcohol. The residue consisting of ditolane-
azotide with a small quantity of lophine was dried, dissolved in the
cold in saturated alcoholic hydrochloric acid, the solution poured into
excess of boiling alcohol, and the whole digested over the water-bath
for some time. Ditolane-azotide was precipitated in minute needles,
whilst lophine and yellow impurities remained in solation. For
analysis, the substance may be recrystallised from boiling benzene,
but, as precipitated, it is practically pure. The yield is about half
the weight of the benzoin employed; the use of sealed tubes is dis-
pensed with, enabling a much larger quantity to be manipulated in a
single operation; and all the troublesome processes of separation
involved in Erdmann’s process are dispensed with.*

* Fusion with ammonium acetate in an open flask may be advantageously sub-
stituted for heating with alcoholic ammonia in sealed tubes in the preparation of
stable compounds usually obtained by the latter process. Thus we find that benzil
gives, by this method, a good yield of benzilam, together with some lophine ; and

830 JAPP AND WILSON ON

We attempted to reduce ditolane-azotide by heating it with
hydriodic acid and amorphous phosphorus, but without success. At
low temperatures it was not attacked; at temperatures above 200°
the nitrogen was eliminated as ammonia,

The granular substance already referred to (p. 826) was finely
powdered, and dissolved by long boiling with alcohol. The solution
was then distilled down to a small volume and allowed to crystallise.
The substance is again deposited very slowly in granular crystals. A
few acicular crystals of ditolane-azotide are generally mixed with it
at this stage; these are mechanically removed as already described
(p. 826). The substance was recrystallised until it presented a per-
fectly homogeneous appearance. It consists of minute but well-defined
oblique tables or prisms, and melts when pure at 199°. It gives no
red coloration with concentrated sulphuric acid. It is somewhat
more soluble in alcohol than ditolane-azotide.

We propose to name the compound benzoinidam.

The analysis of benzoinidam dried at 120° gave the following
results :—

Substance. CO . H,0.
0°1139 0°3458 0°0623
0°1642 0°4978 0°0895

III. 0°1334 gram, burnt with copper oxide in a vacuum, gave
13°79 c.c. of a mixture of nitrogen and nitric oxide, measured dry at
18°8° and under 249°2 mm. pressure. After absorption of the nitric
oxide, there remained 13°79 c.c. of dry nitrogen at 19° and under
246°2 mm. pressure.

IV. 03504 gram gave 22°63 c.c. dry nitrogen + nitric oxide at
15°5° and 3942 mm. After absorption: 22°63 ¢.c. dry nitrogen at
15°7° and 391°1 mm.

V. 0°2666 gram gave 13°79 c.c. dry nitrogen + nitric oxide at
188° and 5003 mm. After absorption: 13°79 c.c. dry nitrogen at
18°8° and 481°7 mm.

These figures do not enable us to decide with certainty as to the
formula of this compound. The formula which agrees best perhaps
with the experimental results is C.,H,,NO, :—

Mr. H. H. Robinson, working in this laboratory about four years ago, ascertained
that phenanthraquinone could in this way be converted in the course of a few
minutes into compounds which Sommaruga obtained by heating it with alcoholic
ammonia in sealed tubes for 36 hours. From the product of the reaction, Mr. Robin-
son isolated diphenanthrylene-azotide and diphenanthrylene-oxytriimide.

AMMONIA-DERIVATIVES OF BENZOIN.

Calculated for Found.

CogH23N Ov. P

Panne ei I. IL. I. —iTV.

Cy... 396 2°96 82:80 82-68 ~- --
Hs... 23 5°68 608 6:06 “= -
N 14 3°46 -- _— 3°96 3°97
O 32 7°90 -- -= -- —

405 100°00

Different preparations were employed in the nitrogen determin-
ations.

The formula C,,H.,NO, is recommended by its simple relation to
that of benzoinam, C,;H.,N,O, the latter compound being a con-
densation-product of 2 mols. of benzoin with 2 mols. of ammonia,
whilst benzvinidam would be formed from 2 mols. of benzoin and
1 mol. of ammonia :—

2C,,H.,0, + NH; = CyH,NO, + 2H,0.
Benzoin. Benzoinidam.

On the other hand, it must be pointed out that the discrepancy
between the calculated and the experimental values for nitrogen
considerably exceeds the experimental error of the very exact method
of determination employed (see Trans., 1886, 475).

We do not, therefore, regard the above formula of benzoinidam as
established.

The smallness of the yield of benzoinidam renders a study of its
decompositions, which would doubtless throw light upon its formula,
a practical impossibility.

We found, however, that the best yield of the compound was ob-
tained by allowing ammonia to act for some weeks upon benzoin in
the cold with exclusion of air. The benzoinidam gradually separates
in moderately large transparent crystals, which may be mechanically
separated from the slender acicular crystals of the other products
present, and purified as already described.

The last mother-liquors from the sealed tubes, after the removal of
the yellow crystalline substance (see p. 826), yielded lophine, identi-
fied by its melting point, Erdmann’s observations as to the formation
of this compound being thus confirmed.

We have to thank Mr. Cosmo Innes Burton for the great care with
which he has made the various analyses given in this paper.

Normal School of Science,
South Kensington.

¢
-
¢
’

832

LXXX.—Note ona Compound fiom Benzil and Isopropyl Alcohol.
By Francis R. Japp, F.B.S., and Juris Rascuen..

Ix a former communication, it was shown (Japp and Owens, Trans.,
1885, 90) that in presence of a small quantity of caustic potash,
benzil and ethyl alcohol unite, with elimination of water, yielding a
compound of the formula C»H.,0,. The reaction occurs according to
the equation: 2C,,H,,0, + C.H;O0 = C»H».O, + H.0:.

We have now made experiments to ascertain whether benzil reacts
in @ similar manner with other alcohols; but only in the case of
isopropyl: alcohol did the alcohol take part in the. reaction. With
propyl alcohol and isobutyl alcohol, there appeared to be merely a
rapid reaction occurring even in the cold; between the bensil and the
potash, to form potassium benzilate.

In the case of isopropyl.alcohol, the experiment was.conducted in
the following manner. The potash (2 grams) was.dissolved in.100 c.c.
of isopropyl alcohol contained in a flask, and to the solution 10 grams
of finely powdered benzil were added, after which the flask was
tightly corked to exclude air, and shaken as long as the benzil
dissolved. After standing for some months, the liquid contained a
quantity of a pulverulent substance in which large transparent yellow
crystals were imbedded. The solid substance was separated by filtra-
tion and washed, first with ether, to remove unaltered benzil, and
afterwards with water,.to dissolve potassium benzilate: The residue
was sparingly soluble in aleohol, but dissolved after long boiling and,
on cooling, was deposited in very lustrous, transparent, faintly
yellowish crystals, of rhombohedral habit. After another recrystal-
lisation from boiling alcohol, the substance was almost colourless.
The crystals contained no alcohol of crystallisation (differing in this
respect. from the ethyl. compound), and melted, with previous
softening, at 147—148°.

The analyses agreed with the formula C3;H2s.0,:—

Substance. GO. H,0.
Di emidion .. 01665. 0°4883 00912
ie éeeneiae 0°1505 0:4411 0°0822
Calculated for Found.
C3,H2,0,.
c _ ~ a II.
Cy 666 ¢.08 372° 80°1 7 79°98 79°93
a 28 6°03 6°09 6°07
ee 64 13°80 —_ a

THORPE AND TUTTON ON PHOSPHORUS TETROXIDE. 833

The benzil and isopropyl alcohol react according to the equation—
2€ 4H 0. + C,;H,0 = C3,H2,0,, + O.

The compound is therefore formed by a process of reduction, and is
not analogous to the ethyl compound,. which, as above stated, is
formed from ethyl. alcohol and benzil. merely by. elimination of
water..

It may be noted that the formula G@,,;H»O, is- homologous with
CyH.0,, the formula ascribed by Limpricht and Schwanert to the
ethyl compound (cf. loc. cit.). We therefore repeated the analysis of
the latter- compound, but merely confirmed the formula Cy»H.,0O,,
assigned to it by Japp and. Owens.

In. giving an account of the above results before the Society
(see Proc., No. 24, p. 204), we stated that we had also obtained a
new condensation compound of. ethyl alcohol with benzoin, formed by
the action of very dilute alcoholic potash upon benzoin in the cold.
We have since found that pure alcoho! does not yield this compound,
and that its formation was due to an impurity contained in the
methylated spirit which we employed, instead of duty-paid alcohol, in "2.
the experiment in question. We therefore reserve an account of thistg
compound until we have more fully ascertained. the conditions of its q
formation.

Normal School of Science,
South Kensington.

JOAIT

WesTTOTNL F

LXXXI.—On: Phosphorus Tetroxide.

By T. E. Tuorrs, F.R.S., and A. E. Turron, Associate of the
Normal School of Science, South Kensington.

Wuen phosphorus glows in air at ordinary temperatures, phosphoric
oxide is the only oxide of phosphorus formed. At slightly higher
temperatures (50—60°), small quantities of phosphorous oxide begin
to make their appearance in addition (see Cowper and Lewes, Trans.,
1884, 10), but it is only when the phosphorus is actually ignited that
any considerable amount of the latter compound is produced.
Blondlot (Compt. rend., 66, 351) has indeed affirmed that the primary
product of the union of phosphorus and oxygen is phosphoric oxide,
and that phosphorous oxide is only formed by the action of the phos-

834 THORPE AND TUTTUN ON PHOSPHORUS TETROXIDE.

phorus upon the higher oxide: 3P,0; + 4P = 5P,0;. At moderately
high temperatures, and with a limited supply of air, there is formed
also, in addition to the phosphorous oxide, a light red or orange
substance, which is frequently mistaken for amorphous phosphorus :
this substance is, however, a true oxide of perfectly definite composi-
tion expressed by the formula P,O. Its nature was first indicated by
Le Verrier (Annalen, 27, 167), who prepared considerable quantities
of it, and studied its reactions. Reinitzer and Goldschmidt (Ber., 13,
845) also obtained the same substance by the action of zinc upon
phosphoryl trichloride at 100°.

It the fumes formed by burning phosphorus in a limited
supply of dry air are aspirated through glass tubes heated by steam,
the deposited oxides are found to be free from admixed phosphorus.
The alleged spontaneous inflammability of phosphorous oxide
is really caused by this admixed phosphorus, which is probably
present in the condition of the white phosphorus described by
Remsen and Keiser (Amer. Chem. Journ., 4, 459). The relative pro-
portion of the phosphorous and phosphoric oxides so formed will,
of course, vary with the amount of oxygen present and the
manner in which the combustion is made. According to Pagels (J. pr.
Chem., 69, 24) the proportion of the phosphorous oxide rarely
exceeds a third of that of the phosphoric oxide. We have, however,
succeeded in getting products in which, as analysis showed, the
amount of the phosphorous oxide was nearly equal to that of the
phosphoric oxide present.

According to Reinitzer (Ber., 14,1884) the mixture of oxides, when
treated with water (care being taken to avoid any great rise of tem-
perature), forms a yellow solution of neutral reaction which coagulates
when heated to 80°. Reinitzer states that this substance is of the
nature of a colloid, since its solution is incapable of passing through
a moistened membrane. We have made various attempts to obtain
this singular body, but without success. On shaking a quantity of
the mixed products of the combustion of phosphorus with pounded
ice, they are rapidly dissolved, occasionally with the formation of an
apparently yellow liquid. This liquid, no matter whether yellow or
colourless, has invariably a strongly acid reaction, and when yellow
could be obtained colourless by repeated filtration. On the filter, a
very small quantity of a light orange extremely finely divided powder
was obtained, which with alkalis and acids gave the reactions of
Le Verrier’s suboxide of phosphorus. The acid solution remained
perfectly clear when heated to 80° with acids, and at once reduced
silver nitrate and mercuric chloride. :

A quantity of the mixed oxides obtained by the slow burning of
phosphorus in air dried by oil of vitriol and phosphoric oxide was

THORPE AND TUTTON ON PHOSPHORUS TETROXIDE. 835

transferred to a tube previously filled with dry carbon dioxide. After
the introduction of the solid oxides, the end of the tube was drawn
out and attached to a Sprengel pump, care being taken in all the
operations to prevent the entrance of moisture.

The tube was carefully exhausted and sealed, and exposed at
gradually increasing temperatures in baths of salts of known melting
points. At about 290°, the white mass in the tube appeared to
change; a considerable quantity of the orange or red suboxide was
observed to form, and at some distance beyond was a sublimate of
clear transparent and highly lustrous crystals. The coloured residue
was subsequently found to be a mixture of phosphoric oxide and
phosphorus suboxide. The crystals could be heated to 100° without
the slightest change ; at this temperature, the edges remained per-
fectly sharp, and no trace of fusion could be perceived. Hence they
were not phosphorus. On heating to about 180° in a sulphuric acid
bath, they volatilised and again formed just above the level of the
liquid in the bath. The sublimate appeared to be perfectly homogeneous
in character; it was made up partly of detached crystals and partly
of aggregates showing precisely the same planes as the separate
erystals. At the first glance the crystals seemed to be cubical in
habitus, but on nearer examination the adjacent faces were found
not to form right angles. Thin crystals gave very marked polarisa-
tion. In a considerable number of the crystals, the edges were modi-
fied, and on rotating under crossed Nicols the maxima of extinction
were found to occur when the modifying edye was parallel with the
plane of vibration of either Nicol, or whenever a crystallographic
axis coincided with a plane of polarisation. The polarisation cha-
racters preclude the possibility of their being tetragonal, and their
whole symmetry indicates that they are not hexagonal. Mr. L. Fletcher,
of the Mineralogical Department of the British Museum, who was
good enough to examine a number of specimens for us, confirms our
opinion that the substance is almost certainly orthorhombic. No
crystallographic measurements are possible in contact with air, as the
crystals are so extremely deliquescent that they seem to liquefy after
a few minutes’ exposure in an open tube. They are almost instantly
dissolved by water with considerable development of heat, forming a
strongly acid solution, unchanged by boiling. With silver nitrate, the
solution gives a white precipitate which rapidly blackens. It also
reduces mercuric chloride to calomel on warming. Potassium per-
manganate solution is, however, only very slowly decolorised by it,
even on warming after the addition of a few drops of sulphuric
acid.

With “ magnesia mixture ” it gives an immediate precipitate of the
double phosphate of magnesia and ammonia. After standing for

§36 THORPE AND TUITON ON PHOSPHORUS TETROXIDE.

some time, the filtered liquid containing excess of magnesium chloride
gave an abundant precipitate of phosphomolybdic acid when treated
with nitric acid and ammonium molybdate solution.

These tests are sufficient to show that the solution contained both
phosphorous and phosphoric acids.

On concentrating the solution in a vacuum over. oil of vitriol, it
forms a colourless viscid syrup, consisting of a mixture of phos-
phorous and phosphoric acids.

On carefully neutralising a moderately concentrated solution of the
mixed acids with soda, no precipitation or formation of a sparingly
soluble salt was observed (wide infra). The neutralised solution placed
in a vacuum over oil of vitriol yielded a thick.syrup, which after long
standing formed a crystalline mass. This behaviour is characteristic
of solutions of neutral sodium phosphite.

The aqueous solution of the soda salts gave the following reactions :—

With silver nitrate,.a yellow. precipitate, changing to brown, and
eventually becoming black. On boiling another portion of the solu-
tion with excess of soda, and again testing with silver nitrate, the
same reaction was observed. Phosphites are not decomposed by
boiling with alkalis, whereas hypophosphites are decomposed into a
phosphate with elimination of hydrogen.

“ Magnesia mixture” gave an immediate precipitate of magnesium
ammonium phosphate. Phosphites and metaphosphates form no pre-
cipitate with magnesia salts in presence of ammonia, unless in very
concentrated solutions. The saline mass consisted therefore of a
mixture of sedium phosphite and orthophosphate.. In order to deter-
mine the composition of the sublimed crystals, weighed quantities of
them were allowed to deliquesce in air, the solution was dilated with
water and evaporated with nitric acid.

Repeated evaporation with nitric acid is necessary in order to
completely oxidise the phosphorous acid. Indeed it is only when the
solution becomes concentrated and the nitric acid is strong that
any oxidation can be perceived. The syrupy liquid from which the
greater part of the nitric acid was expelled by evaporation was diluted
with water and transferred to a platinum crucible containing a known
weight of lime, which before the addition of the phosphoric acid
solution was partially slaked by placing it for about 48 hours in an
atmosphere saturated with water-vapour. The crucible and its
contents were placed on the water-bath and afterwards cautiously
heated to redness until the weight was constant. Preliminary experi-
ments showed that a weighed quantity of lime after partial con-
version into the nitrate can be obtained of the original weight by
ignition. The validity of this method of determining phosphoric
acid in an aqueous or nitric acid solution, which has certain

THORPE AND TUTTON ON PHOSPHORUS TETROXIDE. 837

advantages over the usual process with litharge, was tested by
estimations of the amount of phosphorus in phosphorus trichloride
and in purified amorphous phosphorus. In both cases, numbers were
obtained which were almost identical with those demanded by theory.
Of course care must be taken when weighing the caustic lime that it
does not absorb atmospheric moisture. The simple arrangement
which one of us employed in weighing the hygroscopic titanic oxide
in the course of a series of determinations of the atomic weight of
titanium may be used in weighing the lime (Trans., 1885, 125).
The crucible and its contents are placed under a cover made by
inverting a light beaker with ground edges on a thin piece of sheet-
glass; between the plate and the beaker is a narrow circular strip of
thin sheet caoutchouc against which the beaker is pressed by an
india-rubber band. The cover thus made fits air-tight, and the lime
may be allowed to remain for hours without any increase of weight
being perceived. In order to guard against any change of weight
from the oxidation of the rubber or from its hygroscopic nature, the
cover is weighed immediately after the determination of the weight of
the lime.

Two determinations of the amount of phosphorus in ‘the sublimed
crystals afforded the following results :—

0°5300 gram gave 0°5961 P,O;.
03620 _,,

49°09 per cent. P.

Calculated for P,O,....

P.O; requires 56°36, and P.O, 43°66 per cent. P.

These numbers serve to indicate that the new ‘oxide may be
regarded as phosphorus ‘tetroxide or hypophosphoric oxide, corre-
sponding to the tetroxides of nitrogen and antimony. It is possible
indeed that it is isomorphous with the latter compound, as cervantite,
the native form of antimony tetroxide, is also orthorhombic. Whether,
however, the molecular weight of this compound is expressed by the
formula P,O, remains to be proved. Although, as has been stated, the
substance is volatile in a vacuum, we have not been able to obtain it
in sufficient quantity to determine its vapour-density. A determina-
tion of vapour-density in a vacuum could not be readily carried out;
the extreme rapidity with which the substance deliquesces would also
add to the difficulty. The mode of its behaviour with water finds its
simplest explanation on the assumption that the new oxide is phos-
phoroso-phosphoric oxide, P.O;,P,05, exactly as the tetroxide of anti-
mony is regarded as a compound of antimonious and antimonic oxides.

835 THORPE AND TUTTON ON PHOSPHORUS TETROXIDE.

It may be assamed therefore to be the anhydride of the hypophosphoric
acid or phosphoroso-phosphoric acid discovered by Salzer in the acid
liquid obtained by partially immersing ordinary phosphorus in water
in free contact with air (Annalen, 187, 322). The constitution of this
acid cannot as yet be said to be satisfactorily determined, but from the
general character of its salts, a number of which have been prepared
by Salzer and crystallographically examined by Haushofer and
W. Fresenius, it is most probably represented as P,O,H, or
P,O,(OH)s.

Freshly prepared phosphoroso-phosphoric acid in moderately dilute
aqueous solution does not, however, comport itself like a mixture of
phosphorous and phosphoric acids. It can be boiled without change,
and exerts no reducing action on solutions of the chlorides of mercury,
gold, or platinum, and with silver nitrate it gives a white precipitate
which is not blackened by boiling. Solution of potassium per-
manganate is quickly decolorised by it, especially on warming,
whereas potassium permanganate is but very slowly reduced by
phosphorous acid solution. With sodium carbonate or acetate, it gives
a sparingly soluble acid sodium salt, probably P,O,Na,H.,6H,O;
none of these reactions are given by the solution obtained by the
action of water upon the new oxide. We are disposed therefore to
believe that it is not the anhydride of Salzer’s acid, but that it is the
true phosphorus tetroxide, P,Q,, and that in the mode of its decom-
position with water it behaves like nitrogen tetroxide at low
temperatures :

N,0, + H,O = HNO, + HNO,
P.O, + 3H,O = H;PO,+ H;PO,.

That it is not a mere combination of P.O; and P.O; seems to be
further indicated by the invariable and simultaneous formation in
relatively considerable quantities of the red suboxide. The loose
light powder of the mixed combustion products of the phosphorus
was almost perfectly white before being heated in a vacuum, but after
heating the residue was invariably strongly coloured orange or light
red. That it was the suboxide which was thus formed and not
amorphous phosphorus was established by direct analysis. As the
residue consisted mainly of phosphoric oxide and the red oxide, the
tetroxide would seem to be derived from phosphorous oxide, possibly
in accordance with the equation—

7P,0; = 5P,0, + P,O.

The fact that a crystalline sublimate can be obtained by heating
the combustion products of phosphorus has already been noticed by
Hautefeuille and Perrey (Compt. rend., 99, 33), who, however,

eo = - mew

CHURCH: A CHEMICAL STUDY OF VEGETABLE ALBINISM. 839

describe the substance as one of the three modifications of phosphoric
oxide, which according to them exist. No analytical data are given,
but from the description of the mode of formation of the compound,
its crystalline characters, and the different manner in which it is
deposited according to the conditions of the sublimation, it seems not
improbable that it is identical with the substance described in this
communication.

LXXXII.—A Chemical Study of Vegetable Albinism. Part III.

Experiments with Quercus rubra.

By A. H. Cuurcu, Professor of Chemistry in the Royal Academy
of Arts.

In two former papers (this Journal, 1879, 35, 33—41 ; 1880, 37, 1—6)
I described some conspicuous chemical differences between the white
and the green foliage of seven distinct plants. These differences
were shown in the constituents of the ash, in the nature of the
calcium-compounds present, in the proportion of albuminoid to total
nitrogen, and in the relations of the living foliage to water and to the
atmospheric carbon dioxide. With reference to most of these points,
I am now able to add corroborative data drawn from the examination
of the foliage of a forest tree. Mr. Nicholson, now Curator of the
Royal Gardens at Kew, pointed out to me some time ago the exist-
ence of a very large albino bough on a fine and old specimen of the
scarlet oak (Quercus rubra) of the Eastern United States. This bough
is covered annually with white leaves, barely touched, here and there,
with a few small isolated blotches of green. As is the case with
albino foliage generally, these white oak-leaves are invariably smaller
and thinner than the green leaves of the same age on the other
boughs of the tree. The bough too itself is of less diameter and
length (and so are its branches and twigs) than the green boughs of
the same tree with which it may fairly be compared, the annual
rings being narrower. There is a difference also in the colour of the
bark of the albino bough, for it is a reddish-brown instead of a

purplish-black.

Water, Organic Matter, and Ash.—On the 17th of August, 1880, a
portion of the above-named albino oak bough was sawn off at the
same time as a contiguous green bough; both were growing at a
height of about 20 feet from the ground. The two specimens were
at once taken to the Jodrell Laboratory and submitted to chemical

840 CHURCH: A CHEMICAL STUDY

examination. The two kinds of leaves, after having been separated
from their stems (but retaining their petioles), were analysed with
all the precautions named in my previous papers. They gave the
followmg results, calculated into percentages :—

White. Green.
ME cedtcceneawesos 72°69 58:08
Organic matter........ 24°65 40°33
Bibnssesetardakéseee 2°66 1:59

The contrasts between the white and green leaves shown by these
figures are in entire accordance with those observed in the case of
the plants previously described, and are even more striking. For
example, the difference in the percentages of water here amounts to
no less than 14°61, whilst in my former trials it was as small as 4°83
in one case, that of Plectogyne variegata, and did not, in any instance,
exceed 12°75 (in ivy). Then, again, the ratio between the organic
matter and the ash in the fresh oak-leaves is most characteristic, the
ratio being 100 to 10°8 in the case of the albino foliage, and 100 to
3°9 in the green; the difference between these ratios is larger than
any previously observed.

Composition of the Ash—We may now approach the consideration
of the ash-constituents of these oak-leaves. On account of the
abundance of material, I have been able to make a more complete
examination of this mineral matter than was possible in the case of
the plants before analysed. Potash, lime, magnesia, iron, phosphoric
acid, and silica were the constituents to which my attention was pre-
viously confined ; determinations of manganese, sulphuric acid, and
chlorine have now been added. It may at once be stated that, of the
three last-named constituents, the chlorine alone seems to be decidedly
more characteristic of the albino than of the normal foliage. But
this preponderance of chlorine in the white leaves can scarcely be
due to any special function which it performs in them, but must be
considered to result rather from the much greater quantity of potas-
sium which accompanies it and from the easy diffusibility of the
chloride.

The oak-leaves used in the fuller series of ash analyses were
gathered on the 30th of July, 1886. It had been found that the
first supply, picked on the 17th August, 1880, was not sufficient for
all the determinations required. Still the following percentages
selected from the experiments made with the 1880 crop may be here
given :—

White. Green.
Ash in leaves dried at 100° C. ...... 9°73 3°79
Potash (K,O) in this ash........... 51°28 25°62

Phosphorus pentoxide in this ash.... 1464 17°05

eo 8 & tem m@e a

OF VEGETABLE ALBINISM. 841

Tt will be seen presently that these numbers agree well, so far as
they go, with those obtained with the leaves gathered on tle 30th
July, 1886. The determinations were made in duplicate, and the
following figures are the mean percentages :—

White. Green.

Ash in dry leaves of Quercus rubra .. 833 3°85
100 parts of this ash contained :-—

BUTTE) oeccccassevscecsencs 49°38 29°10
BN CED ev esrr es cwnviereeweres 8°25 24°50
Magnesia (MgO) ...........eeeeeee 6°52 9°55
Ferric oxide (Fe,O;).............06. 0°82 1:24
Manganoso-manganic oxide (Mn,;Q,).. 2°08 2°36
Phosphorus pentoxide (P,O;)........ 14°25 15°80
Sulphur trioxide (SO;) ........... . 718 10°05
GHEE ccccsecesvccevscevsveuse 4°25 1:25
GD v0 as rere crewnwssveerxewsvent 3°15 4°25

95°88 98°10
Deduct oxygen for chlorine ........ 0°96 0°32

94°92 97°78
Carbon dioxide, soda, &c., by difference 5°08 2°22

The above results do little more than confirm the conclusions
reached by means of my inquiries of 1877-79. They show the same
preponderance of potash over lime in the albino foliage, which the
earlier analyses had revealed, of mineral matter over organic, and
also the same remarkable deficiency of lime. As I stated in 1878, so
I may now repeat, and with increased emphasis, “if we determine
the water, the organic substance, and the constituents of the ash, in
each of the following cases, we shall be able to affirm generally, that
white leaves are related to green pretty much as immature leaves are
to mature, tubers to foliage, petals to green bracts, vegetable para-
sites to their hosts.”

Nitrogen-compounds.—The condition, or rather combination, in
which the nitrogen exists in the white leaves as compared with its
state in the green was determined in 1878 in one plant only, Eleagnus
pungens. The same question has now been studied in the case of the
scarlet oak: the results of the two sets of experiments are in satis-
factory accord. In both plants, the total nitrogen and the non-
albuminoid nitrogen are higher in the white foliage than in the
green ; whilst. the content of albuminoid nitrogen is practically the
same in both sorts of leaves, or at least in the dry matter of both
sorts; I reproduce, from. my first paper, the Hlawagnus results,

VO. XLIX. 3 M

842 CHURCH: A CHEMICAL STUDY OF VEGETABLE ALBINISM.

placing them side by side with the corresponding figures of the oak ;
all the numbers represent percentages in the leaves dried at 100°.

Eleagnus pungens. Quercus rubra.

" er aprcine ee
White. Green. White. Green.

Total nitrogen ..........+. 4°23 2°82 3°94 2°78
Albuminoid nitrogen........ 1°83 1°81 2°65 2°41
Non-albuminoid nitrogen.... 2°40 1°01 1:29 0°37

Thus in the two kinds of white leaves the percentage of albuminoid
nitrogen, as compared with the total, is 43 and 67; in the two kinds
of green leaves, 64 and 87. The exact forms of combination in which
the non-albuminoid nitrogen exists in the albino foliage remain to be
ascertained, but there can be no doubt that the excessive quantity of
this nitrogen therein present is to be counted amongst the other
signs of imperfect elaboration which analysis reveals. It should be
stated that the above determinations of albuminoid nitrogen were
made by the phenol method.

Ether Extract, ¥c—I purpose investigating the nature of the sub-
stances extracted by means of absolute ether from the albino oak
foliage: here it must suffice to say that those extractives amounted
to 5°35 and 5°15 parts respectively, from 100 of the dry white and
green leaves. It is remarkable that, notwithstanding the absence of
chlorophyll, the dry matter of the white leaves should yield a rather
larger percentage of ether extract than that of the green; and this
result becomes still more accentuated if we recall the fact that
100 parts of the dry white leaves contained less total organic matter
(91°67 per cent.) than the green (96°15 per cent.). The nature of the
organic substances, other than nitrogen-compounds, and of the matters
soluble in dry ether, also demands investigation. I may, however,
here state that a hot water extract of the white oak-leaves showed,
with iodine solution, the reaction, not of starch, but of erythro-
dextrin, while both white and green leaves contained tannin.

Respiration and Transpiration.—The relation of the albino oak
foliage to carbon dioxide and the atmosphere has been the subject of
a few experiments, the final result of which may be thus expressed :—
1000 square centimetres of surface of these albino leaves (counting
one side only) produced in three hours 7°293 c.c. of carbon dioxide
measured at 0° C. and 760 mm. bar. This figure corresponds with
the elimination of 0°1097 gram of carbon dioxide per 100 grams of
fresh albino leaves. The experiment was made on the 17th August,
1880 ; during two of the three hours of exposure the sun was shining.
The operation was conducted in the way described in my memoir of
1879, but with this difference, that the whole volume of the air in

DITULANE-AZOTIDE AND DIPHENANTHRYLENE-AZOTIDE. 843

contact with the leaves was submitted to analysis. In a similar and
simultaneous experiment with the normal green foliage, the whole of
the carbon dioxide naturally present in the same bulk of air in contact
with the leaves was withdrawn by the same area of leaf; probably
the limits of this absorption were by no means reached.

Some experiments on the aqueous transpiration of the white foliage,
as compared with that of the green, were made on the 17th August,
1880; the results were in but partial accord with those obtained with
holly sprays in 1879 as recorded in my previous paper. Those, how-
ever, were obtained on a dull day, whilst during these more recent
experiments with oak foliage the sun was shining during two hours out
of three. A number of sprays of pure white oak foliage, 24°5 grams
in weight, placed in separate vessels of distilled water, gained
046 gram during the three hours of exposure; whilst 19°5 grams of
the green sprays lost, under the same-conditions, 1°37 grams. When
no water was supplied to the sprays, both the white and the green
foliage lost considerably in weight and very nearly to the same extent.
Further inquiry is needed before any definite conclusions can be
drawn from these results, since, in the previous experiments with
holly, the white sprays not supplied with water lost 6°54 per cent. of
their weight, and the green sprays 10°26 per cent., during 3} hours’
exposure. When water was supplied, the white holly sprays gained
0°29 per cent., and the green holly sprays, instead of losing weight
like the green oak foliage, gained 1°55 per cent.

It remains to be added that the average weights of white and
green oak-leaves, selected so as to be comparable, and deduced from
20 specimens of each sort, were :—

LXXXIII.—Conversion of Ditolane-azotide into Diphenanthrylene-

azotide.

By Francis R. Japp, F.R.S., and Cosmo Inngs Burton, B.Sc.

In a recently published paper on the ammonia-derivatives of benzoin,
it was shown by one of us, in conjunction with Dr. W. H. Wilson,
that the compound to which J. Erdmann assigned the formula
CyHiN (Annalen, 135, 185) has in reality the formula C,,.H»N,. It
was further proposed to change the name of this compound from
benzoinimide—the name given by Erdmann—into ditolane-azotide, in

3 mu 2

844 JAPP AND BURTON: CONVERSION OF DITOLANE-AZOTIDE

order to indicate its analogy to diphenanthrylene-azotide, CosH\.N>.
The view taken as to the relation of these two compounds to one
another was expressed in the following words (p. 828) :—

“In ditolane-azotide two tolane-groups are united by two nitrogen-
atoms; in diphenanthrylene-azotide two phenanthrylene-groups are
united by two nitrogen-atoms. The very great stability of these
compounds renders it probable that the nitrogen-atoms form part of
a closed-chain complex. The constitution of the compounds might,
therefore, be expressed by the following formule :—

C,H;-C—N—C:C,H; C;,H,C—N—C-C,H,
| | > | |
C,H,,;C—N—C-C,H; C;H,C—N—C-:C,H,
Ditolane-azotide. Diphenanthrylene-azotide.*

in which the mode of distribution of the available affinities in the
closed-chain complex is left undecided.”

In the above formule, ditolane-azotide is represented as containing
four phenyl-groups, diphenanthrylene-azotide as containing two di-
phenylene-groups. Now Graebe has shown that compounds containing
two phenyl-groups may frequently be converted, with elimination of
hydrogen, into the corresponding diphenylene-compounds by passing

them in a state of vapour through a red-hot tube: for example,
stilbene, when thus treated, yields phenanthrene; diphenylamine
yields carbazole. It seemed, therefore, possible that. in this way
ditolane-azotide might be converted into diphenanthrylene-azotide, a
reaction which would exactly correspond with Graebe’s synthesis of
phenanthrene from stilbene. Such a transformation would go far
towards proving the correctness of the assumption made in the paper
just referred to, as to the analogy of ditolane-azotide to diphenanthry-
lene-azotide.

We, therefore, distilled a quantity of ditolane-azotide through a
red-hot tube filled with fragments of glass; but from the tarry
distillate nothing definite, beyond unchanged ditolane-azotide, could
be obtained.

On again reading Erdmann’s paper, however, we found an observa-
tion which appeared to point to an actual transformation of ditolane-
azotide into diphenanthrylene-azotide. Erdmann mentions that in
heating the former compound with soda-lime, he obtained a yellow,
crystalline sublimate ‘“‘ which, by its behaviour towards concentrated
sulphuric acid, showed itself to be a mixture of at least two com-
pounds. One portion of the crystals dissolved in this acid with a

* Sommaruga, who first gave a name to this compound, calls it diphenanthrene-
azotide, which is very much as if ethylene dichloride were to be termed ethane-
dichloride.

INTO DIPHENANTHRYLENE-AZOTIDE. 845

magnificent deep-blue colour, another portion with a beantiful violet-
red colour” (Annalen, 135,185). This deep-blue colour with con-
centrated sulphuric acid appeared to us to indicate the presence of
diphenanthrylene-azotide, of which this is the most characteristic
reaction; whilst the violet-red colour would be due to a mixture of
this compound with unchanged ditolane-azotide, the latter compound
yielding with sulphuric acid a blood-red colour, which, when mixed
with the blue from the diphenanthrylene-azotide, might be expected
to give a violet.

In order to test the correctness of this assumption, we mixed
5 grams of ditolane-azotide with powdered soda-lime and distilled it
from combustion tubing over granulated soda-lime heated to medium
redness. This time there was no tar formed; but a sublimate of flat
yellow needles collected close to the red-hot portion of the tube—a
behaviour highly characteristic of diphenanthrylene-azotide. Further
along the tube, where the temperature was somewhat lower, a pale-
yellow sublimate collected ; whilst in a test-tube placed over the open
end of the combustion tube the sublimate was white. The yellow
needles gave, with concentrated sulphuric acid, an almost pure blue
colour; the pale-yellow sublimate gave a violet, in which, while
dissolving, separate streaks of red and blue were visible; whereas
the white sublimate gave a pure red colour, and consisted of unaltered
ditolane-azotide. In order to remove ditolane-azotide, the product of
the distillation was boiled with benzene, which dissolves this com-
pound readily, whilst it only takes up traces of diphenanthrylene-
azotide. The residue was then treated with a concentrated solution
of hydrochloric acid in alcohol, which also dissolves ditolane-azotide,
but not diphenanthrylene-azotide. This last treatment was further
intended to remove any traces of lime that might have been mechani-
cally carried over during the distillation. Finally, the undissolved
portion was washed with boiling alcohol, and dried for analysis. Thus
purified, it consisted of a yellow, glistening crystalline powder. The
yield was about one-tenth of the weight of ditolane-azotide taken.

Before analysing the substance, various qualitative tests were
applied, in which it was compared with diphenanthrylene-azotide
obtained from phenanthraquinone. The identity of behaviour in the
two substances was absolute. Heated between watch-glasses, both
sublimed at a very high temperature, condensing in flat, iridescent,
yellow needles, sometimes obliquely truncated, sometimes notched at the
end as if by twinning. Both dissolved in sulphuric acid, giving deep-
blue solutions, indistinguishable from each other in tint. On gradually
diluting the solutions with water the colour changed in both cases to
orange, and on further dilution a practically colourless, amorphous
substance was precipitated. According to Sommaraga diphenanthry-

846 DOBBIN AND MASSON: ACTION OF THE

lene-azotide does not melt at the boiling point of sulphur. A portion
of the substance prepared from ditolane-azotide, enclosed in a capillary
tube and heated in sulphur vapour, showed no sign of fusion.

Analysis of the product gave figures agreeing with those required
for diphenanthrylene-azotide, C.,H,.N2 :—

Substance. CO. H,0.
cecccece 0°1729 0°5593 0°0720

II. 0°1370 gram, burnt with copper oxide in a vacuum, gave
13°79 c.c. of a mixture of nitrogen and nitric oxide, measured dry at
17°5° and under 502°7 mm. pressure. After absorption of the nitric
oxide there remained 13°79 c.c. of dry nitrogen at 17°5° and under
460°2 mm. pressure.

Calculated for Found.
C23H¢No. ceemmemen:.
A I II.

.
Che 000 336 88°42 88°22 —
err 16 4°21 4°63 —
Bigee soa 28 7°37 — 7°53

380 100°00

There is no doubt, therefore, that the compound obtained by the
distillation of ditolane-azotide with soda-lime is diphenanthrylene-
azotide, and it thus appears that J. Erdmann unconsciously synthesised
this complex phenanthrene-derivative in 1865—seven years before the
discovery of phenanthrene itself.

Normal School of Science,
South Kensington.

LXXXIV.— Action of the Halogens on the Salts of Organic Bases.
Part II. Tetramethylammonium Salts.

By Leonard Dossy, Ph.D., University of Edinburgh, and Orme
Masson, M.A., D.Sc., Professor of Chemistry in the University of
Melbourne.

In a previous paper (Trans., 1885, 56), we detailed the results of
experiments on the action of the halegens on the salts of trimethyl-
sulphine. In that paper we showed that by acting with the halogens
on the haloid trimethylsulphine salts (Me,SX), a series of addition
products could be obtained, of the general formula Me,SX;, the

HALOGENS ON THE SALTS OF ORGANIC BASES. 847

various members of which differed very much amongst themselves in
their chemical and physical properties. Only in the cases where the
products contained iodine, were we able to get crystalline substances
fit for analysis ; and those containing more than one atom of iodine in
their molecules were either dark crystals which rapidly lost iodine
when exposed to the air, or merely formed a dark tarry mass of poly-
iodide, from which a crystalline product could not be separated.
Those products which did not contain iodine, but only bromine or
chlorine or both, appeared as orange-red or yellow, more or less
viscous liquids which decomposed slowly in dry air, losing halogen;
or more rapidly in contact with water, trimethylsulphine salts
remaining behind.

Although the peculiar character of some of these substances did not
allow of their composition being accurately determined, it was shown
that it might be expressed by the general formula Me,SX, (X = I, Br,
or Cl), and that all the compounds suggested by this formula may be
obtained synthetically.

The property of taking up iodine, bromine, or chlorine to form
addition products is by no means confined to the salts of trimethyl-
sulphine, but seems to be common to the salts of many organic bases.
We find that the reactions take place in the case of the salts of the
nitrogen and of the phosphorus bases at least as readily as (and with
those of the phosphorus bases perhaps more readily than) with the
trimethylsulphine salts. It is the addition products obtained from
the tetramethylammonium salts that we wish to describe in the

present paper.

I. Action of Iodine on Tetramethylammonium Iodide.

This action has been described by Weltzien (Annalen, 99, 1), who
employed an alcoholic solution of iodine, and a solution of the iodide
in aqueous alcohol, and obtained what he says was a pentiodide, or
a mixture of this with triiodide, according to the proportion of iodine
added. He describes these two compounds at considerable length,
and mentions many of their chemical properties, but he does not give
any analytical data except one iodine estimation in the triiodide.

Ii. Action of Iodine on Tetramethylammonium Bromide and Chloride.

On shaking up the dry solids with iodine dissolved in alcohol,
solutions were obtained which, on evaporation in a vacuum, yielded
small slightly-coloured crystals. The crystals from the bromide were
orange, only slightly deliquescent in air, and rapidly lost iodine,
leaving bromide. The crystals from the chloride were pale-yellow,

848 DOBBIN AND MASSON: ACTION OF THE

and also lost iodine rapidly, leaving behind a very deliquescent
residue of chloride.

III. Action of Bromine on Tetramethylammonium Iodide.

The crystals of the iodide are turned brown directly they come in
contact with bromine-vapour. When liquid bromine touches the dry
iodide, a hissing sound is heard and much heat is developed. In the
preparation of the addition product, bromine was added until there
appeared to be no further action. The product is a dark pasty mass,
which, in a current of air, loses the excess of bromine and becomes
a hard solid of a colour much resembling that of cinnabar. This
solid is reduced to powder, washed with ether, and dissolved in hot
absolute alcohol, in which it is fairly soluble. On cooling, the alco-
holic solution deposits feathery crystals of a deep orange colour. The
mother-liquor, on evaporation, gives a crop of considerably larger
crystals, mixed with colourless, slightly deliquescent crystals of a
substance which proved to be tetramethylammonium bromide.

The portion of the substance used for analysis was part of the crop
of crystals deposited from hot alcohol, and was dried for about 14
days over strong sulphuric acid.

The numbers obtained by analysis prove the substance to be tetra-
methylammonium dibromiodide, Me,NI,Br,, and are as follow :—

I. 0°3243 gram gave 0°1597 gram CO, and 0°1075 gram H,0.
II. 0°3114 " 0°1529 » » O1016 ~

Found.
Calculated. a In
13°43 13°39
3°68 3°62

A nitrogen determination was not made in this case, but the nitro-
gen was estimated in the analogous dichloriodide, to be described
further on.

Tetramethylammonium dibromiodide crystals are quite stable, and
not deliquescent even in moist air; they are almost insoluble in and
seem to be scarcely decomposed by water. They are sparingly
soluble in cold absolute alcohol, but moderately soluble in hot, and
from the hot solution an abundant crop of crystals is deposited on
cooling. They melt at 190° with very little apparent decomposition,
their decomposition with evolution of iodine-vapour taking place at a
considerably higher temperature. Treated with ammonia solution,
they behave very like the corresponding trimethylsulphine compound,
forming a dark, almost black substance which explodes on drying.

HALOGENS ON THE SALTS OF ORGANIC BASES. 849

Action of Gaseous Ammonia on Tetramethylammonium Dibromiodide.

Having observed that trimethylsulphine dibromiodide united with
dry.ammonia to form a definite compound, Me,SIBr.,2NH;, we tried
the action of gaseous ammonia on the tetramethylammonium com-
pound, and found that here also an analogous substance was formed.
There was no apparent change in the orange crystals at first, nor until
the ammonia had been passed for about an hour, when it was noticed
that the orange colour had changed to a paler yellowish colour.
Later, the substance became almost colourless, but minute yellow or
orange particles could still be distinguished when constant weight
had been reached.

0°3616 gram dibromiodide became 0°3960 gram by absorbing dry

ammonia.
Gain per cent.

On exposing the product to the air for some days, the ammonia was
given off again pretty rapidly, and the original orange colour of the
dibromiodide was quite restored.

IV. Action of Chlorine on Tetramethylammonium Iodide.

A quantity of the finely powdered iodide was placed in a small
flask, and a current of dry chlorine passed over it. The powder,
which at first became dark, very soon afterwards assumed a pale-
yellowish colour, and then no further change appeared to take place.
As the action was manifestly not finished, the heat of a steam-bath
was applied, and after a few minutes there was a violent reaction.
In certain patches the powder became black and semi-fluid. A species
of ebullition, accompanied by the formation in considerable quantity
of whitish fumes took place, extending outwards from these patches,
and eventually spreading throughout the greater part of the
substance. As the wave of ebullition passed outwards, the substance
resolidified at each centre into a hard black solid, which, by the further
action of chlorine, became light-yellow in colour to a certain extent,
but complete conversion of the whole into a homogeneous yellow
substance did not take place until the mass had been reduced to
powder, and this had again been heated in the current of chlorine
for about half an hour. The yellow powder obtained was dissolved,
like the bromine compound, in hot alcohol, and the solution, on
cooling, yielded yellow feathery crystals which were washed with
alcohol and dried over strong sulphuric acid. The mother-liquor on
evaporation also yielded some slightly larger crystals.

850 DOBBIN AND MASSON: ACTION OF THE

Analysis gave the following results, which agree with the composi-
tion Me,NICI, :—

I. 0:3042 gram gave 0°1930 gram CO, and 0°1234 gram H,0.
II. 0°3170 - 02047 —,, " 0°1302 -

III. 0°2785 gram was ignited with soda-lime, and the products of
combustion were passed through dilute hydrochloric acid. With
platinic chloride, this acid gave a precipitate containing 0°0971 gram
platinum.

Found.

Caleulated.

Tetramethylammonium dichloriodide, like the dibromiodide, is quite
stable in air, but is more easily decomposed on heating, and the
alcoholic solution appears to undergo a good deal of change on
long keeping, or when heated for some time. The crystals do not
melt until a temperature of 216° to 220° is reached, and not without
partial decomposition. A deep brown liquid is formed, and a slight
yellowish distillate.

Ammonia solution acts on the dichloriodide as it does on the di-
bromiodide, forming a black explosive substance, possibly iodide of
nitrogen.

Dry ammonia did not act on the dichloriodide nearly so readily, nor
to such an extent as on the dibromiodide. As soon as it came in con-
tact with the gas, the dichloriodide at once turned a rather dark
olive-green colour, which very soon became lighter and continued to
get lighter for a considerable time. After half an hour, the whole
had assumed a greyish-green colour which did not undergo further
change. The least trace of moisture caused this substance to turn
superficially a colour exactly resembling cocoa. The absorption of
ammonia was very slow from the beginning, and after three and a
half hours (when it had all but ceased) the percentage gained was
only 7°4, whereas the formula Me,NIC1,,2NH; requires a percentage
gain of 12°5. This reaction proving so sluggish, it was not further
investigated.

A compound was obtained by Weltzien (Annalen, 99, 11) by the
recrystallisation from water of Me,NICl;, which had the composition
of tetramethylammonium dichloriodide, and seems to be identical with
it, although the analytical results given do not agree very closely
with theory. He obtained the substance Me,NICI; by passing

HALOGENS ON THE SALTS OF ORGANIC BASES. 851

chlorine in excess into a solution of tetramethylammonium iodide,
and evaporating the liquid produced.

Tilden (this Journal, 1866, 145) obtained tetrethylammonium di-
chloriodide by the action of protochloride of iodine on a hydrochloric
acid solution of tetrethylammoniam chloride. This substance ap-
pears to have properties resembling those of the corresponding
methyl compound.

In one of his original papers describing the properties of tetrethyl-
ammonium compounds, Hofmann mentions haiogen compounds which
probably resembled those obtained by us from tetramethylammonium
salts, although he thought them to be substitution compounds. He
says (Roy. Soc. Trans., 1851, 370) :—“ The action of various chemical
agents upon the compounds of tetrethylammonium gives rise to a series
of very remarkable substances. Chlorine, bromine, and iodine convert
the base into substitution products, in which the basic character of
the original atom has disappeared; of these the bromine compound
is distinguished by its splendid appearance, crystallising as it does
from alcohol, in long, magnificent orange-yellow needles. The
iodine compound too is very beautiful; it forms either on addition of
iodine solution to the base, or on evaporating a solution of the iodide
exposed to the action of the air. In fact, it is difficult to avoid the
formation of this substance in recrystallising iodide of tetrethyl-
ammonium.”

Although a future contribution detailing the properties of these
bodies is promised, we have not been able to find their description in
any subsequent papers.

V. Action of Iodine Monochloride on Tetramethylammonium Bromide.

This action very much resembles that of bromine on the iodide,
in the phenomena displayed, the liquid chloride producing a hissing
sound when poured on to the dried and powdered bromide. The result-
ing solid was well washed with ether, and recrystallised from hot
vicohol. The recrystallised substance very much resembled in colour
the dichloriodide, and analysis proved it to be this, contrary to our
expectations, as we had fully expected to get the chlorbromiodide,
Me,NIBrCl.

0°3358 gram gave on combustion 0°2152 gram CO, and 0°1532 gram

Calculated for Calculated for
Me,NICly. Found. Me,NIBrCl.

17°48 151
5°07 3°8

From this analysis it would appear that in the case of tetramethyl-

852 DOBBIN AND MASSON: ACTION OF THE

ammonium bromide (unlike that of trimethylsulphine bromide), the
bromine is displaced by another halogen (presumably chlorine) when
acted on by iodine monochloride.

VI. Action of Bromine on Tetramethylammonium Bromide and
Chloride.

The dry solids were both rapidly turned orange-brown by bromine-
vapour. Both slowly became semi-liquid, and then solidified again
somewhat on the surface. After this stage no further change was
observable. On passing dry air for some time over the products,
both became semi-liquid again, giving off bromine rapidly, and the
dark-brown liquids, on continuing to pass air, partially resolidified.
The semi-solid substances were treated.with ether, when they imme-
diately solidified. The solids were washed repeatedly with ether, then
dissolved in alcohol and evaporated in a vacuum, when yellowish
crystals were obtained in each case. These compounds were both very
unstable, losing bromine quickly on exposure to moist air, and
becoming colourless. The colourless chloride residue deliquesced
very rapidly, and that from the bromide less rapidly. The crystals
of both addition products were darkened by ammonia solution, and
the dark substances produced dissolved afterwards in the ammonia
solution with evolution of small bubbles of gas—probably nitrogen.

VII. Action of Chlorine on Tetramethylammonium Bromide and
Chloride.

The dry solids when placed in a current of dry chlorine, were only
slowly affected by it, and very little change of colour was noticeable.
Both became slightly yellow in colour, but no tendency to become
liquid was observed. After the gas had been passed for a con-
siderable time, the products were examined. Both gave off a little
chlorine in the air, but not at all abundantly. Moistened with water,
no formation of bubbles of gas was observable in the case of the
bromide, but there was a distinct darkening of colour, as if a displace-
ment of bromine by chlorine only took place in presence of water. In
the case of the chloride there was a slight evolution of chlorine. Both
products, when exposed to moist air, quickly deliquesced—that
obtained from the bromide far more rapidly than the almost non-
deliquescent bromide itself would have done, showing that it had been
converted by the chlorine in part at least into the very deliquescent
chloride.

The rate of absorption of chlorine by the chloride was very slow,
as the following quantitative experiment shows :—

HALOGENS ON THE SALTS OF ORGANIC BASES. 853

Weight of Time Increase of Increase

substance taken. in chlorine. weight. per cent.
0°5942 1} hours 0-0928 15°6
oa 0-1058 178

The calculated percentage increase for Me,NCI; is 64°8.

VIII. Action of Bromine on Tetramethylammonium Sulphate.

The sulphate, dried for many hours at 100° in a current of dry air,
was placed in a current of air charged with bromine-vapour. It
was at once coloured brown, and took up bromine very rapidly, but
did not become liquid except to a slight extent.

A quantitative experiment was made in this case, of which the
details are given below :—

Weight of sub- Time in bromine- Increase of Increase
stance taken. vapour. weight. per cent.
0°3364 gram 23 hours 0°7134 gram 212
a 0:7194 ,, 213
Time over solid Increase reduced Increase
caustic potash. to: per cent.
1 day 0°3286 gram 97°6
3 days 0:2909 ,, 80°6
17 ,, 0:2867__,, 80°4

The percentage increase for Br, (calculated for [MeN ].S0,,Br.) is
73'7, so the result of our experiment cannot be said to give much
definite information. The sulphate, in an atmosphere of bromine, is
capable of taking it up in large quantity, and the product, in dry air,
loses. a. large proportion of this bromine again with great rapidity,
to a certain point, when further loss practically ceases.

IX. Action of Chlorine on Tetramethylammonium Sulphate.

The perfectly dry colourless sulphate soon became pale-yellow in a
current of chlorine, but it did not become liquid, nor was any further
change visible after the first 10 minutes or so. The product placed
in air rapidly deliquesced, giving off a little chlorine. Mixed with
water, it dissolved very rapidly, small bubbles of chlorine being given
off for some time.

Before leaving the experimental portion of this paper, we wish to
mention an experiment made with tetrethylphosphonium iodide, which
was the only suitable phosphonium compound at our immediate
disposal.

854 DOBBIN AND MASSON: ACTION OF THE

X. Action of Bromine on Tetrethylphosphonium Iodide.

Bromine acts at once on the dry solid, turning it reddish-brown,
and, after a time, a dark brown liquid is formed from the melting of
the product. This liquid crystallises when the excess of bromine goes
off, needle-like crystals shooting from the edges towards the centre.
The solid thus formed dissolves pretty readily in and crystallises from
cold alcohol. The crystals were bright and shining, and of an orange-
yellow colour. We have not in the meantime further examined

them,

The close analogy between the tetramethylammonium-derivatives
described in the foregoing section and the trimethylsulphine-derivatives
described in our former paper (already referred to) is obvious; the
method of formation being identical in the two cases, and the chemical
and physical properties strikingly similar. Various facts already
established by others have convinced us, moreover, that the salts of
many other organic bases are capable of yielding analogous compounds
with the halogens; and preliminary experiments of our own have
shown that, as might be expected, the reaction is also applicable to
salts of phosphorus bases.

In evidence of the former statement, we may mention the papers of
several investigators.

In a paper on the “ Polybromides of the Tetrammonium Bases,”
Marquart (J. pr. Chem., 1, 429) describes, with some detail, the
formation and properties of tetrethylammonium tribromide. To
prepare this substance, he simply mixed tetrethylammonium bromide
with bromine-water. It was fairly stable, and the analytical results
are moderately close to theory. He says, that in trying to prepare
the corresponding methyl compound he found it to be decidedly less
stable, being decomposed on recrystallisation, with formation of
bromide as well as tribromide. An attempt to get a polychloride by
acting with chlorine-water on the chloride failed, as of course it
must fail, polychlorides being decomposed in presence of water.

Addition products have been obtained from pyridine by the aetion
of halogens. Hofmann (Ber., 12, 984) describes dibromopyridine—a
substance which crystallises in fine orange crystals. Ramsay (see
Watts’ Dict., 3rd Supp., 1700) acted on picoline and its salts with
bromine, &c., and obtained various analogous products.

Ostermayer has lately published several papers on the “ Action of
Iodine Monochloride on Alkaloids,” &c. (Abstr., 1855, 672; Ber., 18,
591 and 2298), to which we will only refer here.

M. Dittmar has also recently published a paper on the “ Reaction
between Iodine Monochloride and the Alkaloids ” (Ber., 18, 1612). In

HALOGENS ON THE SALTS OF ORGANIC BASES, 855

it he makes two statements on which we wish to comment. He says
(p. 1614) that all alkaloids which react with iodine monochloride,
producing the characteristic bright- yellow precipitates with the
ammonia reaction, to be described farther on, contain one or more
pyridine-groups ; and the number of the halogen-groups added on,
corresponds, in general, with the number of pyridine-groups contained
in the particular vegetable base. The ammonia reaction which he
describes is the formation of a dark-green substance from the iodine
monochloride compounds by the action of ammonia solution.

Farther on (p. 1616), he says that this ammonia reaction gives a
key to the constitution of the substance which yields it, and that it is
an excellent reagent for the group of chlorine and iodine, this group
alone giving it.

We have shown, both here and in our previous paper, that the
formation of a yellow crystalline compound with iodine monochloride
takes place with substances which do not contain any pyridine-group,
and that the ammonia reaction takes place with substances which do
not contain both chlorine and iodine, e.g., tetramethylammonium and
trimethylsulphine dibromiodides. The fallacy of the first of these
statements has been pointed out already by Ostermayer (Ber., 18, 2299),
who shows that a yellow precipitate, giving the ammonia reaction, can
be readily obtained from concentrated solutions of caffeine.

The reaction between the halogens and iodine monochloride, and
organic bases and their salts, is a very general one, and applicable to
nitrogen, sulphur, and phosphorus compounds alike ; there is, moreover,
little doubt but that similar addition products are formed in the case
of all those bases, the radicles of which are sufficiently positive,* and
which are not completely broken up by the action of the halogens.

The exact constitution of these addition products is a matter of
some interest, and has a bearing on the vexed question of the
existence or non-existence of molecular compounds as distinct from
atomic compounds. Of the trihaloid derivatives of trimethylsulphine,
which may be taken as representative of the whole class, some would
undoubtedly, if their properties were studied apart from those of the
rest, be classed as molecular compounds by the supporters of the
theory of two distinct kinds of combination; whilst others would,
with almost equal certainty, be considered as true atomic compounds,
on account of their stability when heated, and their other properties.
Of the series-—Me,SIBr., Me,SIBrCl, Me,SICl,, Me,SBr,, Me;SBr,Cl,
Me,SBrCl,, Me,;SCl;, the higher members—those which contain an
atom of iodine—are crystalline substances, which remain solid and
unchanged at temperatures near 100°, and, when they do decompose,

* Compare the action of halogens on diethylenesulphide-methylsulphine iodide
(Masson, Trans., 1886, 239).

856 DOBBIN AND MASSON: ACTION OF THE HALOGENS.

do not show any tendency to split into their original components. In
this respect they do not in the least resemble hydrated salts or other
typical so-called molecular compounds.* The lower members, on the
other hand, though formed in a precisely analogous manner, display a
tendency—which becomes more and more marked as we descend in the
series—to split again into free halogen and trimethylsulphine salt ; and
they do so, in fact, more or less rapidly at the ordinary temperature
unless preserved in an atmosphere of halogen vapour.

We thus have a continuous series of substances, to whose members
it would be absurd to attribute totally different constitutions, but
which shows the atomic compound characters at one end and the
molecular compound characters at the other. This fact appears to us
to give some support to the position of those who deny the existence
of two modes of combination, distinct in kind from each other.

Dr. Armstrong, in his address to the Chemistry Section of the
British Association (see Reports, 1885), suggested that the so-called
molecular compounds are, in reality, atomic compounds which owe
their existence to the special attraction of negative atoms for negative
atoms, and their capability of manifesting towards one another a
certain “residual affinity” which they do not manifest towards
positive atoms.

Whichever of the two possible atomic constitutions be attributed to
the compounds we are treating of, it appears to us that they afford
excellent examples of this accumulation of negative atoms, a process
which would appear to be manifested in a still higher degree by the
unstable, but definite, compounds which they themselves form with
ammonia, e.g., Me;SIBr,,2NH;. In the one case, if the constitution

be represented by the general formula Me,sS—X< = it is merely the

halogen that is affected; in the other, MeSEX, it is the central nega-
tive atom (sulphur, nitrogen, or phosphorus, as the case may be) that
manifests residual affinity towards the halogens. On the whole we
incline towards the latter view, for, as we have already pointed out,
the sulphates of trimethylsulphine and tetramethylammonium form
unstable compounds with bromine and chlorine, which bear the closest
resemblance to the corresponding trichlorides. A closer study of the
phosphonium addition products than we have yet made may throw
some further light on the question, for these should be more stable
than the others (on account of the greater affinity of phosphorus for
the halogens) if our view be correct.

* Compare on this point Kolbe (remark on Marquart’s paper, J. pr. Chem., 1,
435).

RENNIE: GLYCYPHYLLIN. 857

We have again to express our thanks to Professor Crum Brown for
his advice and assistance during the progress of our work.

LXXXV.—Glycyphyllin, the Sweet Principle of Smilax glycyphylla.

By Epwarp H. Rewyniz, M.A. (Sydney), D.Sc. (London), Professor
of Chemistry in the University of Adelaide, South Australia.

In a paper by Dr. C. R. Alder-Wright and the author which appeared
in the Journal of this Society (Trans., 1881, 237), a partial descrip-
tion was given of a crystalline substance extracted from the leaves of
Smilax glycyphylla,* a plant growing in abundance on the shores of
Port Jackson, and common on the coasts of the northern parts of
New South Wales and the southern parts of Queensland. The inves-
tigation was not completed from want of material, and although the
writer obtained a grant from the Society some three years ago for a
further examination of this substance, other duties have prevented,
until quite recently, the completion of the work.

The results obtained confirm in some respects those published
in the paper above referred to, but have shown that the formula
therein proposed needs alteration. In brief, the substance has the
formula C2,H.,0, + 3H,O when crystallised from aqueous ether, and
C.,H,O, + 44H,0 when crystallised from water. When boiled with
dilute sulphuric acid, it undergoes decomposition in accordance with
the equation—

CHO, + 2H,0 = C5H,,05 + CsHy0,,

forming phloretin and isodulcite, and is therefore closely allied to
phlorizin. A discussion of these results will be found in the detailed
account which follows.

Extraction and Purification.

In order to obtain the material for this investigation, about 75 lbs.
weight of the leaves and stems were macerated for several days with
strong spirit (till the leaves became tasteless) and the alcohol dis-
tilled off. The dark-coloured syrupy liquid was then exhausted with
ether (a tedious process, as ether does not dissolve the substance very

* As the leaves treated were not submitted to an expert, it is possible that among
them were some from Smilax australis, but the majority were from the above-
mentioned species.

VOL. XLIX. 3.N

858 ‘RENNIE: GLYCYPHYLLIN,

freely), and the ether distilled off. The residue always solidified on
cooling to a dark-brown or yellow mass of crystals mixed with more or
less semi-fluid resinous matter. The best method of effecting a rough
purification was found to be solution and repeated crystallisation
from hot water eontaining a little alcohol, rejecting any resinous
matter which refused to dissolve easily. By this means, the greater
part of the colouring matter could be removed and a mass of yellowish
crystals was obtained, which, however, obstinately refused to become
colourless by crystallisation alone. The further purification was
effected by one of two methods.

1. The substance was dissolved in warm water, allowed to cool until
just about to crystallise, then quickly shaken up with ether, and the
latter rapidly poured off through a dry filter into a dry flask. On
standing, a great part of the substance separated out as a perfectly
white mass of minute crystals.

2. The substance was dissolved in hot water, some acetate of lead
added, and the precipitate, which formed in larger or smaller quantity
according as the solution was darker or lighter in colour, filtered
through a hot-water filter and rejected. Sulphuretted hydrogen was
then passed through the solution, the lead sulphide filtered off, the
liquid being still warm enough to keep the substance in solution, and
the filtrate allowed to crystallise. It was sometimes found necessary
to repeat this process in order to get a perfectly white product, but
it was finally adopted as the easier.

Composition and Properties.

The crystals deposited from a warm aqueous solution consisted of
very brittle, slender four-sided prisms, often from one-half to one inch
in length, if the solution had been slowly cooled. Whether separated
from ether (under the conditions described above) or from water, they
contain water of crystallisation, but not the same in amount (details
given below). The followingare the results of several combustions
of different samples of anhydrous substance (dried at from 100—110°),
Nos. I, II, III prepared by the ether process, and the remainder by the
acetate of lead process :—

Substance. CQ . H,0.
B osseee 0°3535 0°7735 0°1980
DE soaeee 0°2680 0°5865 0°1425
| ae 0°3073 0°6689 0°1640
BY sweses 0°3390 0°7375 0°1825
wT cesses 0°3255 0°7100 0°1755
,) oe 0°3249 0°7085 0°1735
WEE sévcoe 0°2902 0°6352 0°1550

ececee 0°2474 0°5404 0°1314

THE SWEET PRINCIPLE.OF SMILAX.GLYCYPHYLLA.. 8099

Calculated for -

Co, Hoy.
paceman linemen, I. i. TI. IV.
Cy.... 252 60°00 59°67 59°35 59°36 5933.
Hy... 24 5°70 6°22 5°83 5°93 5°98
QO,.... 144 34°30 oe os a= —-
420 100-00
v. Vi. Vil. VIII.
ee 59°49 59°47 59°69 59°57
Mewes 5°99 5°93 5°93 5°90

D secs — — — —

The following table gives in percentages the results of several water
determinations made in different samples, dried at 100—110° :—

Mean.

Crystallised from ether 11°31 11°88 11°84 11:43 11:13 11°54
Crystallised from water 1628 1617 1631 1611 — 16°22

The formula C,,H,,0, + 3H,O requires 11°39 per cent. of water.
‘ C..H,O,+44H,0 , 1616 ,, o

A combustion of an air-dried specimen (crystallised from water)
gave carbon 49°85 per cent., hydrogen 6°55 per cent., theory requiring
carbon 50°29 per cent., and hydrogen 6°58 per cent., thus confirming
the above results.

For the present at least, 1 propose to retain for this substance the
name glycyphyllin, suggested in the joint. paper already quoted, as it
appears to be quite different from the substance described under the
name of smilacin and from other known substances. It has no well-
defined melting point, but at temperatures varying from about 110°
to 115° it begins to show signs of change, and as the temperature
rises, is gradually transformed into a caramel-like substance, which at
175—180° melts with decomposition. It is only very sparingly soluble
in cold water, but sufficiently so to communicate to the solution its cha-
racteristic strong liquorice-like taste. It is easily soluble in hot water
and alcohol, soluble to some extent in ether, insoluble in chloroform,
benzene, and light petroleum. It dissolves in solutions of the caustic
alkalis, and the solution gradually assumes a rich reddish-brown
colour on exposure to the air. It does not reduce Fehling’s solution.
Its solution is not precipitated by ordinary acetate of lead, but is
immediately precipitated by the basic acetate.

It was not possible in this case to get an accurate estimate of the
yield of material, but it could not have been more than one-half, and
was probably nearer one-third per cent.

860 RENNIE: GLYCYPHYLLIN,

Decomposition by Dilute Sulphuric Acid.

When an aqueous solution of glycyphyllin is boiled in a flask with
reversed condenser, with the addition of a small quantity of dilute
sulphuric acid, the liquid soon becomes turbid, and rapidly deposits
an almost colourless crystalline powder. On cooling, a small quantity
more separates out, the liquid having at the same time acquired
the power of reducing Fehling’s solution. As the precipitated sub-
stance is never obtained quite white, it is purified by dissolving
it in a hot mixture of alcohol and water (about equal volumes),
adding acetate of lead, filtering off any slight precipitate, passing
sulphuretted hydrogen, &c., &c. The filtrate, on cooling, deposits a
snow-white crystalline mass, which is anhydrous. The following are
the analytical numbers obtained from different samples :—

Substance. CO . H,0.

D ésbeee 0°2030 0°4857 0°1015

rr 02690 06486 0°1295

Ben wesees 0-2885 0°6962 0°1370

rer 0°2558 0°6170 0°1220

, eee 02535 0°6063 0°1225

ee 0°2532 0°6064 0°1244

Calculated for
C,;H,405.

ceo I II. ITI. IV. es Vi.

Cis.. 180 65°69 65°25 65°76* 65°82* 65°78* 65°22 65°31
Hy.. 14 511 552 5°35 5°28 5°29 536 0-540
O;.. 80 29°20 — — — — — —_

274 100-00

These numbers agree satisfactorily with those required by the
formula C,;H,O;, which is ascribed to phloretin, and, on further
examination, leaves no doubt as to the identity of the two substances.
The substance obtained as described above is almost insoluble in cold
water, and only very sparingly soluble in boiling water, from which
it crystallises on cooling in minute shining prisms. It is easily
soluble in alcohol, and crystallises from a hot mixture of alcohol and
water in groups of needles radiating from a centre. It is soluble
also in ether and hot glacial acetic acid. It melts at 250° with decom-
position. Stas gives the melting point of phloretin as 180°, and this
at first misled me, but it has siuce been shown by Schiff (Ber., 14,

* In these combustions, tle oxygen which had been used in the combustions was
inadvertently allowed to remain in the tubes while weighing, thus accounting to
some extent for a somewhat high percentage of carbon.

alt

THE SWEET PRINCIPLE OF SMILAX GLYCYPHYLLA. 861

303) that the melting point is close to 250°. It dissolves with yellow
coloration in solutions of the caustic alkalis and of the alkaline
carbonates, and is reprecipitated on acidifying. The solution in
strong ammonia, after a short time, deposits golden-yellow scales.
When boiled with strong caustic potash for from 15 to 20 minutes,
the solution is no longer precipitated on addition of an acid, and is
found to contain phloroglucol and an acid which is to all appearance
identical with phloretic acid.

To isolate these substances, the alkaline solution was first slightly
acidified with hydrochloric acid, rendered slightly alkaline with sodium
hydrogen carbonate, and then exhausted with ether. The residue
after the evaporation of the ether was decolorised partly by the aid of
animal charcoal and partly by the acetate of lead process, and then
crystallised several times from water. The crystals contained water of
crystallisation, and had a very sweet taste. The melting point of the
anhydrous substance was found to be 208—209°.* The aqueous
solution gave a deep reddish colour with ferric chloride. On com-
bustion, the following numbers were obtained :—

Substance. CO. H,0 lost.
Dédseaess 02220 0°4627 —
BP eccaccas 0°3046 0°6328 0°1357
Calculated for
Ce H,03
| I II
C, 72 57°14 56°84 56°66
Bihasane 6 4°76 _ 495
hisses 48 38°10 — —
126 =100°00

There can be no doubt, therefore, that the substance in hand was
phloroglucol.

The solution from which the phloroglucol had been removed was
acidified by hydrochloric acid, and again exhausted with ether. On
evaporation, a mass of dark-coloured prismatic crystals remained.
These were decolorised by animal charcoal and several times recrystal-
lised from hot water; by slow cooling, crystals were often obtained an
inch in length. On combustion, the following numbers were ob-
tained :—

Substance. CO. H,0.
D eepuedes 0°2532 06000 0°1440

* This is the melting point given in all the more recent accounts of phloroglucol,
although in Miller’s Chemistry and elsewhere it is stated to be about 220°.

862 RENNIE: GLYCYPHYLLIN,

Calculated for
C,H 903.
ota z
Teh ok sdakerewcs 108 65°09 64°62
Dn seteneseen 10 6°03 6°32
Miiecsesccnonws 48 28°88 —
166 100 00

The silver salt was prepared by precipitating a solution of silver
nitrate with a solution of the barium salt (obtained by boiling the
acid with pure barium carbonate), washing the precipitate in a
partially darkened room, drying it in the dark over sulphuric acid,
and finally at 100°. It gave the following numbers on analysis :—

Substance. CO). H,0. Ag.
DP éeakvass 0°3086 0°4494 00995 01212
Calculated for:
CyH,AgO3.

| L
er 108 39°56 39°72
Piessticadtencs 9) 3°58 3°29
a 108. 39°56 39-27
isda seeneees 48 17°30 —

273 100°00

The acid melted at 127—128°, and its aqueous solution gave a
slight violet coloration with ferric chloride. In other physical
characteristics, it agreed closely with the description given of phloretic
acid, and it is obviously identical with the acid previously ob-
tained by fusing glycyphyllin with caustic potash at 250°. Trinius,
who has synthesised phloretic acid (Annalen, 22'7, 262), states that
when pure it gives no coloration with ferric chloride, but says that a
sample prepared from phlorizin gave this coloration, and he concludes
that the reaction is due to a small quantity of impurity in the acid,
and that Rochleder’s so-called isophloretic acid was really pure
phloretic acid. The acid prepared as above described, even when
repeatedly crystallised from water, converted into the silver salt, and
regenerated from the latter, still gave the reaction. The colour
produced, however, could hardly be said to be well marked, and is
probably due, as Trinius suggests, to the presence of some impurity
(phloroglucol ?) which obstinately adheres to the acid prepared in this
way. It may be remarked, however, that a single drop of ferric
chloride solution, such as is usually employed as a reagent, gives
scarcely a perceptible coloration, but the colour deepens considerably

THE SWEET PRINCIPLE OF SMILAX GLYCYPHYLLA. 863

on the addition of a further quantity. This may perhaps account to
some extent for the discrepancy.

Examination of the Filtrate after Separation of the Phloretin.

The filtrate from the phloretin was neutralised by the addition of
barium carbonate, the barium sulphate and excess of carbonate filtered
off, and the-solution evaporated to dryness. During this process, more
or less of a red colouring matter was produced, but the syrup which
remained:soon crystallised. The crystals were washed once or twice
with a little cold strong alcohol which removed most of the colouring
matter, and then repeatedly crystallised from ordinary alcohol, but
the product so obtained always had a slight molasses-like odour, and
on combustion gave numbers which did not agree well with any
probable formula. It was, therefore, recrystallised two or three times
from very strong alcohol (98) per cent.) until quite white and
free from: odour. Specimens so prepared melted at 93—94° when
slowly heated in a capillary tube. It gave up one molecule of water
on being heated at 100°. Three different samples gave the following
results on ‘analysis :—

Substance. CO.. H,0.

Mewsceesiee 0°2513 0°3657 0°1800

Pbssoesees 0:2495 0°3609 0°1785
eee 0°2660 0°3828 0°1849

Calculated for
CH, 40¢.
een I. IT. III,

Te sas0 72 39°55 39°68 =. 39°45 39°25
Hy.... 14 7°69 7°95 7°95 7°72

OQ, .... 96 52°76 — — —

182 100°00

This substance was found to be-sparingly soluble in cold alcohol, but
easily soluble in the boiling liquid, from which, or from water by slow
evaporation, it separated in well-formed monoclinic crystals agreeing
exactly in appearance with the crystals of isodulcite obtained by Will
from naringin, and figured by him (Ber., 18, 1318). It reduced
Fehling’s solution on boiling. The quantity obtained was insufficient
to determine its specific rotatory power, but the above evidence is, I
think, sufficient to identify it as isodulcite.

864 RENNIE: GLYCYPHYLLIN,

Quantitative Examination of the Decomposition by Dilute Sulphuric
Acid.

In order to determine the equation representing the decomposition
of glycyphyllin by dilute sulphuric acid, weighed quantities were
boiled with suitable proportions of. the dilute acid in a flask connected
with a reflux condenser until the decomposition was judged to be
complete. The liquid was allowed to cool, the phloretin filtered off,
well washed with cold water, washed off the filter with alcohol, and
the alcoholic solution evaporated todryness. The filtrate and washings
were then extracted with ether to remove small quantities of phloretin
which still remained in solution, and the residue from the evaporation
of the ether added to the main portion, the total product being finally
dried at 100—110°. Four experiments of this kind were made with
the following results :—

Glyeyphyllin

(anhydrous). Phloretin. Per cent.
BE sscccces 1°224 0°8055 65°63
TET wccccces 11°479 7°463 65°01
|) ae 5°239 3°474 66°31
TV cccccess 5°0033 3°317 66°30

In the last of these experiments, the liquid freed from phloretin
was made up to 1 litre, and titrated with Fehling’s solution. It was
found that 26 c.c. were required to reduce 10 c.c. of Fehling’s solution
(1 e.c. = 0°005 glucose). Assuming that 10 c.c. Fehling’s solution
are reduced by 0°0549 gram isodulcite,* as recently determined by
Will (loc. cit.), we have—

1000 x 0°0549
26

This is equal to 42°2 per cent. on the glycyphyllin taken. Con-
sidering the formation of small quantities of secondary products, such
as colouring matter, these numbers agree as well as can be expected
with those indicated by the equation—

CHO, + 2H,0 = C,;H,,0; + CsHuO.,

which requires 65°22 per cent. of phloretin and 43°33 per cent. of
isodulcite.

total isodulcite = = 27115 grams.

Discussion of above Results.

The formation of isodulcite by the decomposition of glycyphyllin

* In an actual experiment, a number very close to this was obtained for the
reducing power of isodulcite, but the details and therefore the exact numbers were
unfortunately lost.

THE SWEET PRINCIPLE OF SMILAX GLYCYPHYLLA. 865

is interesting as affording another instance of a so-called glucoside
which yields isodulcite and not glucose on boiling with dilute acids.
It has been shown that isodulcite has more than once been mistaken
for glucose (Ber., 15, 215; 18, 1316, et seg.). During the course of
this investigation, it seemed possible that this might have happened
in the examination of phlorizin, in which case glycyphyllin might be
identical with that substance. The close resemblance in physical
properties seemed to render this probable. Thus Schiff (Ber., 14, 303)
states that phlorizin “ melts at 110° to a semi-fluid opaque mass, which
on further heatine became again solid and porcelain-like, and at
170—171° melts to a clear yellow fiuid;” this is very like the
behaviour of glycyphyllin. Schiff also calls attention to Loewe’s
criticism on the results published with reference to phlorizin, especially
as regards the discrepancies in the analytical numbers obtained.

On closer examination, however, the differences appear to be too
great to be accounted for by any experimental error. The highest
percentage of carbon obtained in the analysis of phlorizin is about
2 per cent. lower than that required for glycyphyllin; the amount of
water of crystallisation found in phlorizin by several experimenters
corresponds with only 2 mols. H,O; and, finally, although Hesse
states (Miller’s Organic Chemistry, 5th edition, p. 582) that the sugar
obtained from phlorizin is essentially different from glucose, yet he
finds for its specific rotatory power in a solution freshly prepared
[«]p = 89, a number more than 10 times as great as that given by
several observers for isodulcite.

Ihave sent to Trommsdorf for a supply of phlorizin, in order to
compare it with glycyphyllin, and if possible to ascertain the relation-
ship between these two substances, and I hope, ere long, to communi-
cate the results to the Society.

VOL. XLIX. 80

INDEX OF AUTHORS’ NAMES.

TRANSACTIONS. 1886.

Abrahall, J. L. H., phenylsulphonic
anhydride, 692.

Allen, F., obituary notice of, 342.

Andrews, T., obituary notice of, 342.

Anschiitz, R., and P. N. Evans, con-
tributions to our knowledge of the
chlorides of antimony, 708.

Anschiitz, R., and J. W. Leather,
Pipitzahoic acid, I, 709.

Armstrong, H. E., the theory of the
interaction of carbon monoxide, water,
and oxygen gases, 112.

Armstrong, H. T., and A. K. Miller,
the decomposition and genesis of
hydrocarbons at high temperatures.
1. The products of the manufacture
of gas from petroleum, 74.

Aston, Miss E. See Pickering, 8.
U.

Bailey, G. H., notes on an analysis of
koppite, 153.

— on a method of separation and
estimation of zirconium, 149, 481.

Bayley, T., on the analysis of alloys
and minerals containing the heavy
metals, selenium, tellurium, &c., 735.

Bellenot, G. See Perkin, W. H.,
jun.

Brierley, J.T., on some new vanadium
compounds, 30.

the electrolytic preparation of
vanadious sulphate, 822.

Brothers, H. E. See Smith, W.

Brown, A. J., on an acetic ferment
which forms cellulose, 482.

the chemical action of pure culti-
vations of Bacterium aceti, 172.

Brown, E. O., obituary notice of, 344,

Burton, 0.1. See Japp, F. BR.

Calman, A. See Perkin, W.H., jun.

Carnelley, T., and J. Schleselman,
amidodiphenylsul phonic acid and azo-
dyes from diphenyl, 380.

Church, A. H., a chemical study of
vegetable albinism. Part III. Ex-
periments with Quercus rubra, 839.

Clements, G. W. H., obituary notice
of, 345.

Coutts, J.F.H. See Smith, W.

Dechan, M., detection and estimation
of iodine, bromine, and chlorine,
682.

Divers, E., and T. Shimidzu, mer-
cury sulphites and the constitution of
sulphites, 533.

Dixon, H., the combustion of carbonic
oxide and hydrogen. I. The action of
steam on carbonic oxide, 94.

—— the combustion of cyanogen, 384.

Dobbin, L., and O. Masson, the
action of halogens on the salts of
organic bases. Part II. Tetramethyl-
ammonium salts, 846.

Fehling, H. v., obituary notice of, 346.

Field, F., obituary notice of, 347.

Fries, H. H., contributions to a know-
ledge of cyanuric derivatives, 314.

—— further contributions to the know-
ledge of cyanuric chloride and other
cyanuric derivatives, 739.

Friswell, R. J.,and A. G. Green, the
constitution of diazobenzeneanilide
and its relatior to amidoazobenzene,
746.

868
G.

Gladstone, J. H., on essential oils.
Ill. Their —— refractive and dis-
rsive energy, 609.
ofsdetoee *y . and A. Tribe,
aluminium BP III. Aiuminium
orthocresylate and its products of
decomposition by heat, 25.

Graebe, C., and A. Rée, some com-
pounds obtained by the aid of 8-sul-
phophthalic acid, 522.

Grif iths, A.B., on the usé of ferrous
sulphate in agriculture, 114.

H.
Higgin, J. Obituary notice of, 351.

James, J. W., action of phosphorus
pentachloride on ethylic diethylaceto-
acetate, 50.

—— derivatives of taurine, 486.

Japp, F. R., and C. I. Burton, con-
version of ditolane-azotide into di-
phenanthrylene-azotide, 843.

Japp, F. R.. and J. Raschen, note on
a compound from benzil and isopropyl
alcohol, 832.

on the action of phosphoric
sulphide on benzonephenone, 478.

Japp, F. R., and W. H. Wilson, on
ammonia-derivatives of benzoin, 825.

Japp, F. R., and W. P. Wynne, on
imabenzil, 473.

on the action of aldehydes
and ammonia on benzil, 462.

Jordan, A. E.,and T. Turner, on the
condition of silicon in pig-iron, 215.

—_—

Klein, E., bacteriological research from
a biologist’s point of view, 197.

Kohn, C. A., some ammonium com-
pounds and other derivatives of a-1’
hydroxyquinoline, 500.

Laurie, A. P., on the measurement of
the electromotive forces produced by
the combination of cadmium and
iodine in the presence of water, 700.

INDEX OF AUTHORS.

McGowan, G., some derivatives of
thiocarbamide, 190.

McLeod, H., on the electrolysis of
aqueous solutions of sulphuric acid,
with special reference to the forms of
oxygen obtained, 591.

Masson, O., on sulphine salts contain-
ing the ethylene radicle. I. Diethyl-
ene-sulphide-methyl-sulphine salts,
233.

on sulphine salts containing the
ethylene radicle. II. Debn’s reaction
between ethylene bromide and ethyl
sulphide, 249.

Masson, O. See Dobbin, L.

Meldola, R., and F. W. Streatfeild,
a method of investigating the con-
stitution of azo- and diazo-derivatives
and analogous compounds, 624.

Miller, A. K. See Armstrong,
H. E.

Munro, J. H. M., the formation and
destruction of nitrates and nitrites in
artificial solutions, and in river and
weil waters, 632.

Nicol, W. W. J., water of crystallisa-
tion, 690.

O’Sullivan, C., on the presence of
“ yatlinose’’ in barley, 70.

— on the sugars of some cerenls and
of germinated grain, 58.

Perkin, W. H., sen., on the magnetic
rotation of mixtures of water with
some of the acids of the fatty series,
with alcohol, and with sulphuric acid ;
and observations on water of crystal-
lisation, 777.

—— on the constitution of undecylenic
acid as indicated by its magnetic rota-
tion, and on the magnetic rotation of
mono- and di-allylacetie acids, and of
ethy] diallylmalonate, 205.

the formation of acids from alde-

hydes by the action of anhydrides

and their salts and of ketones from
the compounds resulting from the

union of anhydrides with salts, 317.

INDEX OF AUTHORS. 869

Perkin, W. H., jun., and G. Bellenot,
paranitrobenzoylacetic acid and some
of its derivatives, 440.

Perkin, W. H., jun., and A. Calman,
benzoylacetic acid and some of its
derivatives, IV, 154.

Pickering, S. U., modification of
double sulphates, I, 1, 12.

on water of crystallisation, 411.

the influence of temperature on
the heat of chemical combination,
260.

Pickering, S. U., and Miss E. Aston,
on multiple sulphates, 123.

Ramsay, W., and S. Young, eva-
poration and dissociation, a study of
the thermal properties of acetic acid,
790.

note on the vapour-densities

of chloral ethyl-alcoholate, 685.

—— on the vapour-pressures of
bromine and iodine, and on iodine
monochloride, 453.

—— —— on the vapour-pressures of
mercury, 37.

Raschen,J. See Japp, F. R.

Rée, A., 8-sulphophthalic acid, 510.

Rée, A. See also Graebe, C.

Rennie, E. H., “glycophyllin,” the
sweet principle of Smilax glycy-
phylla, .

parabenzylphenol and its deriva-
tives, III, and on an isomeric benzyl-
phenol, 406.

Richardson, A., determination of
vapour-pressures of alcohols and
org ‘nic acids, and the relations exist-
ing between the vapour-pressures of
the alcohols and organic acids, 761.

Rideai, S., noteon the action of ammo-
nia on chromy] dichloride, 367.

Ruttan, R. F., trimethyldiethylamido-
benzene, 813.

Schleselman, J. See Carnelley, T.

Senier, A., contributions to the history
of cyanuric chloride and cyanuric
acid, 311.

— on the action of hexabromacetone
on urea, 693, 743.

Shimidzu, T. See Divers, E.

Smith, W., Coutts, J. F. H., and H.
E. Brothers, examination of the

phenol constituents of blast furnace
tar obtained at the Gartsherrie iron-
works, 17.

Snape, H. L., certain aromatic cyanates
and carbamates, 254.

Stallard, G., the monobromophthalic
acids, 187.

Stern, A. L., on the action of bromine
on phosphorus trichloride, 815.

Stuart, C. M., action of cinnamic and
salicylic aldehydes on malonic acid,
365,

Streatfeild, F. W. See Meldola,
R

Stuart, C. M., the relation of benzal-
malonic acid to its mononitro-deiiva-
tives, 357.

Thorpe, T. E., and A. Tutton, phos-
phorus tetroxide, 833.

Tribe, A., obituary notice of, 352.

Tribe, A. See also Gladstone, J. H.

Turner, T., the influence of remelting
on the properties of cast iron; notes
on Sir .W. Fairbairn’s 1853 experi-
ments, 493.

—— the influence of silicon on the pro-
perties of cast iron, ITI, 130.

Turner, T. See Jordan, A. E,

Vv.

Veley, V. H., some sulphur compounds
of barium, 369.

Watts, F., on the essential oil of lime
leaves (Citrus limetta), preliminary
notice, 316.

Williams, Miss K. J., and W. Ram-
say, the estimation of free oxygen
in water, 751.

Williams, W. C., reactions supposed
to yield nitroxyl or nitryl chloride,
222.

Wilson, W. H. See Japp, F. R.

Witt, O. N., the eurhodines, a new
class of colouring matters, 391.

Wynne, W.P. SeeJdapp, F. R.

Young,S. See Ramsay, W.
302

TRANSACTIONS.

INDEX OF SUBJECTS.

1886.

Ammonium salts, nitrification of, 643,
654.

8-sulphophthalimide, 519.

-—— thiocyanate, nitrification of, by
soil, 637.

Anhydrotaurine, 490.

Anilylmelamine, 743.

Annual general meeting, March 30th,
1886, 329.

Anthraquinone, m-chloro-, 531.

Antimony pentachloride, 708.

Ash of white and green leaves of Quercus
rubra, 839.

Azo- and diazo-derivatives, &c., method
of investigating the constitution of,
624

Azo-dyes from diphenyl, 380.

Bacillus anthracis, 200.

tuberculosus, 201.

Bacteriological research from a biologist’s
point of view, 197.

Bacterium aceti, action of, on alcohol,
178.

— — action of, on cane-sugar, 181.

— — action of, on carbohydrates,
179.

— —— action of, on dextrose, 179.

— —- action of, on mannitol, 182.

chemical action of pure culti-
vations of, 173.

Balance sheet of the Chemical Society,
from March 23, 1885, to March 22,
1886, 355.

of the Research Fund, from
March 23, 1885, to March 22, 1886,
356.

Barley, sugars of, 58.

—— presence of raffinose in, 70.

Barium hydrosulphide, 370, 375, 376.

Barium hydroxyhydrosulphide, 369.

monosulphide, 369, 370.

sulphur compounds of, 369.

thiocarbonate, 370, 378.

Benzalmalonic acid and its mononitro-
derivatives, 357.
tion of, by water, 357.

Benzamide, action of cyanuric chloride
on, 312.

Benzenyldicinnylenediamine, 469.

Benzil, action of aldehydes and ammo-
nia on, 462.

and isopropyl alcohol, compound
from, 832.

Benzoin, ammonia-derivatives of, 825.

Benzoinam, 825.

Benzoinidam, 830.

Benzoinimide, 825, 828.

Benzophenone, action of phosphoric
sulphide on, 478.

Benzophenonidene
481.

Benzoylacetic acid and its derivatives,
154.
440.

Benzoylbenzoic acid, m-chloro-, 530.

Benzoy] chloride, action of silver cyan-
urate on, 313.

Benzoyldiazoamidobenzene, 749.

Benzylphenol, an isomeric, 406.

p-Benzylphenol and its derivatives, 406.

Bromine, detection and estimation of,
682.

—— melting and boiling points, 454.

vapour-pressures of, 455.

Butyl phenyl ketone, 161.

Butyrone, preparation by
method, 322.

decompozi-

pyrothiophosphite,

p-nitro-, and its derivatives,

Perkin’s

Cadmium and iodine, measurement of

INDEX OF SUBJECTS.

the electromotive forces produced by
the combination of, in presence of
water, 700.

Caoutchene, refraction and dispersion
equivalents of, 618.

Carbohydrates, action of Bacterium
aceti on, 179.

Carbonic acid, action of hydrogen on,
in contact with red-hot platinum,
101.

action of hydrogen on, in pre-

sence of anhydrous phosphoric acid,

102.

and hydrogen, action of in-

duction sparks on, 104.

monoxide, water and oxygen gases,
theory of the interaction of, a note
on Mr. H. B. Dixon’s paper on the
action of carbonic oxide on steam,
112.

—— oxide, action of steam on, in con-
tact with red-hot platinum, 97.

action of steam on, in pre-

sence of potash, 101.

-—— and hydrogen, combustion

of, 94.

and steam, action of induc-

tion sparks on, 103.

incomplete combustion of, in
presence of varying quantities of
steam, 104, 109.

Cedrenes, refraction and dispersion equi-

- valents of, 617.

Cellulose, an acetic ferment which forms,
432.

Cerealose, 73.

Cereals, some, sugars of, 58.

Chloral ethyl-alcoholate, vapour-densi-
ties of, 685.

Chlorine, action of, on nitric peroxide,
226.

--— detection and estimation of, 682.

Chromyl dichloride, action of ammonia
on, 367.

Cinnamaldehyde, action of, on malonic
acid, 365.

Cinnamic acid, experiments on the
preparation of, by Perkin’s method,
320.

m-nitro-a-bromo-, 361.

Cinnidimabenzil. 472.

Cinnimabenzil, 471.

Citrenes, refraction and dispersion equi-
valents of, 615.

Citrus limetta,
leaves of, 316.

Coal-tar, blast-furnace, an examination
of the phenol constituents of, ob-
tained by the Alexander and McCosh
process at the Gartsherrie ironworks,
Part 1, 17.

essential oil of the

871

Copper sulphate, anhydrous, heat of
dissolution of, 294.

monohydrated, heat of dis-

solution of, 310.

pentahydrated, heat of dis-
solution of, 296.

Coumarincarboxylic acid, 367.

Cyanogen, combustion of, 384.

Cy — acids, supposed isomeric, 693,

43.

— chloride, 739.

action of benzamide on, 312.
action of sodium formate,

benzoate, and acetate on, 311, 312.

derivatives, 739.

o-Cresyl ether, 25.

—

—-

Denitrification, 667, 681.

—— and nitrification, alternate, 669.

Dextrose, action of Bacterium aceti on,
179.

liallylacetic acid, magnetic rotation of,
212.

Diallylmalonic acid, magnetic rotation
of, 211.

Diazoamidobenzene, p-dinitro-, 626.

Diazobenzeneanilide, constitution of,
and its relation to amidoazobenzene,
II, 746.

Diazobenzenebenzylanilide, 749.

Diazobenzenemethylanilide, 748.

Diazoethylamidobenzene, p-dinitro-, 630.

Dibenzoyldicinnylenediamine, 469.

Diethyldiphenylenedicarbamate, 256.

Diethylene disulphide, constitution of,
234.

preparation of, 235.
Diethylene-sulphide-methyl-sulphine
hydroxide, 247.
Diethylene-sulphide-methyl-sulphine
salts, 237.
Dimethoxy-diethyl-acetone, 57.
Dimethylene-ethane, and its preparation
and oxidation, 81.
Di-a-naphthylamidocyanuric
315.
Di-8-naphthylamidocyanuric chloride,
740

chloride,

Diphenanthrylazotide, conversion of di-
tolaneazotide into, 843.

Diphenyl diphenylenedicarbamate, 256.

Diphenylene diisocyanate, 255.

Diphenylene-diurethane, 256.

Diphenylfurfurandicarboxylic acid, 168.

Diphenylhydrazinecyanuric chloride,
742.

Diphenylsulphonic acid, p-amido-, 380.

303

872

Diphenyltoluylene dicarbamate, 258.
Dithiocarbamide dichloride, 191.

—— diiodide, 195.

dinitrate, 194.

Ditolaneazotide, 829.

conversion of, into diphenanthryl-
azotide, 843.

Ditoluylenediamidocyanuric chloride,
741.
Double sulphates, modifications of,

Parts I and II; Part II. Specific
heat determinations, 1, 12.

Electromotive forces produced in the
reaction of cadmium, iodine, and
water, 700.

Essential oils, ITI, their specific refrac-
tive and dispersive energy, 609.

Ethanesulphonimide, 490.

Ethoxyhydrodiethylquinolium hydrox-
ide, 505.

Ethoxyhydroethylquinoline, dinitro-,
509.

ethiodide, 505.

Ethy] alcohol, action of Bacterium aceti
on, 175.

hydrated, magnetic rotation
of, 780.

vapour-pressures of, 762, 768,
771, 773.

allylacetate, magnetic rotation of,
213.

allyl-p-nitrobenzoylacetate, 451.
—— B-chlorophthalate, 529.
diallylmalonate, magnetic rotation
of, 209.

diethylacetoacetate, action of am-
monia on, 58.

action of phosphorus penta-
chloride on, 50.
diethylchloracetoacetate, 52.

— action of sodium methylate
on, 54.

diethyldichloracetoacetate, 54.
action of sodium methylate

on, 56.
—— dimethoxydiethylacetoacetate, 57.
— di-,-nitrobenzoylsuccinate, 452.
—— diphenylfur!urandicarboxylate,
167.
— ethyl-p-nitrobenzoylacetate, 450.
isobut, lbenzoylacetate, 165.
isopropy!|benzoylacetate, 164.
—— methylbenzoylacetate, 156.
a-methyl-8-chlorocinnamate, 158.
—— methoxydiethylacetoacetate, 55.
—— p-nitrobenzoylacetate, 447.
— - p-vitrophenylnitrosamine, 631.

INDEX OF SUBJECTS.

Ethyl nitroso-p-nitrobenzoylacetate, 449.

—— propylbenzoylacetate, 160.

a-propyl-8-chlorocinnamate, 162.

sulphide, action of ethylene bro-
mide on, 249.

—— undecylenate, 206.

—— —— magnetic rotation of, 207.

Ethylamine, nitrification of, by soil,
633.

Ethylene bromide,
sulphide on, 249.

Eurhodine, 394.

Eurhodines, a new class of colouring
matters, 391.

Eurhodol, 397.

Evaporation and dissociation: IV, a
study of the thermal properties of
acetic acid, 790.

action of ethyl

Ferment, acetic, which forms cellulose,
432.

Ferrous sulphate, antiseptic properties
of, and its action on vegetable para-
sites, 119.

—— experiments with, on rose

trees, 122.

method of applying, to the

land, 122.

retentive properties of, for

ammonia aud phosphoric acid, 121. .

use of, in agriculture, 114.

Fluorescein, 8-chloro-, 530.

Formic acid, hydrated, magnetic rota-
tion of, 778.

vapour-pressures of, 765, 774,

776.
Fulminates, constitution of, 582.

Gas, the products of the manufacture
of, from petroleum, 74.

Gelatin, nitrification of, by soil, 641.

Glycerol, vapour-pressures of, 764.

Glycyphyllin, 857.

Grain, germinated, sugars of, 58.

Graphite in pig iron, 220.

Heat of chemical combination, influence
of temperature on, 260.

—— of formation of double salts, 287.

of hydration of salts, 417.

Hydrates in solution, 275.

Hydrogen, action of, on oxygen, 107.

INDEX OF SUBJECTS.

Hydrogen and carbonic acid, action of
induction sparks on, 104.

—— and carbonic oxide, combustion of,
94.

Hydrocarbons, decomposition and gene-
sis of, at high temperatures: I, the
products of the manufacture of gas
from petroleum, 74.

Hydrochloric acid, action of oxygen on,
under the influence of light, 608.

Hydroxyhydroethylenequinoline, 508.

Hydroxyhydroethylquinoline ethiodide,
505.

Hydroxyhydromethylquinoline meth-
iodide, 51.

benzil chloride, 506.

8-Hydroxyphthalic acid, 522.

8-Hydroxyphthalide, 525.

8-Hydroxyphthalimide, 524.

Hydroxypipitzahoic acid, 728.

—— dibromide, 732.

a-1’-Hydroxyquinoline and its deriva-
tives, 500.

Hydroxythymoquinoneimide, 725.

Hypomercurosic sulphite, 567.

Imabenzil, 473.

Iodine, detection and estimation of, 682.

melting and boiling points of, 454.

monochloride, 461.

vapour-pressures of, 458.

Iron, cast, influence of remelting on the
properties of, 493.

—_— influence of silicon on the
properties of, 130.

—— pig, condition of silicon in, 215.

—— graphite in, 220.

Isoamyl alcohol, vapour-pressures of,
764, 771, 773.

—— phenyl ketone, 166.

Isobutyl alcohol, vapour-pressures of,
763, 771, 773.

—— phenyl ketone, 165.

Isobutyldiphenylglyoxaline, 467.

Isobutyric acid, vapour-pressures of,
766, 774, 776. ‘

Isoprene, refraction and dispersion
equivalents of, 619.

Isopropyl alcohol and benzil, compound
from, 832.

Isovaleric acid, vapour-pressures of,

767, 774, 776.

Ketones, formation of, from the com-
pounds resulting from the union of
anhydrides and salts, 317.

873

Koppite, analysis of, 153.

Lead nitrate, action of phosphorus oxy-
chloride on, 224.

Leaves, white and green of Quercus
rubra, analyses of, 839.

‘Levulose, formation of, from mannitol,

184.
Lime-leaves, essential oils of, 316.
Lithium sulphate, anhydrous, heat of
dissolution of, 309.
—— monohydrated, heat of disso-
lution of, 310.

Magnesium sulphate, anhydrous, heat
of dissolution of, 291.

heptahydrated, heat of disso-
lution of, 292.

Magnetic rotation of water with some
of the acids of the fatty series, with
an and with sulphuric acid,

- 777.

Malonie acid, action of cinnamic and
salicylic aldehydes on, 365.

Mannitol, action of Bacterium aceti on,
182.

formation of levulose from, 184.

Mercuric hydrogen sulphite, 554.

sodium sulphite, 538.

oxysulphite, 546.

—— sulphite, “ normal,” 535.

Mercurosic sulphite, 535, 559.

Mercurous anhydrosulphite, 566.

Mercury sulphites and the constitution
of sulphites, 533. Oy

constitution of, 574.

reduction of, with sulphurous
acid, 575.

—— vapour-pressures of, 37.

Methoxyhydrodimethylquinolium _hy-
droxide, 504.

Methoxyhydromethylquinoline, 501.

—-— methiodide, 503.

Methoxymethyl butyl ketone, 55.

Methoxymethyl-ethyl-acetone, 55.

Methyl alcohol, action of Bacterium
aceti on, 177.

vapour-pressures of, 762,
768, 771, 773.

—— benzyl-p-nitrobenzoylacetate, 446.

—— £-chlorophthalate, 529.

—— f-hydroxyphthalate, 524.

— p-nitrobenzoylacetate, 444.

_— ketone from Citrus limetta,

317.

a-Methy]-8-chlorocinnamic acid, 158.

Methyldiazoamidobenzene, 748.

Methyldiphenylglyoxaline, 465.

a-Methyl-8-phenylhydroxypropionic
acid, 159.

ern y-dibromo-, 189.

a-Naphthylamidocyanuric chloride, 314.

8-Naphthylamidocyanuric chloride, 740.

Naphthylenetoluquinoxaline, amido-,
400

8-Naphthylmelamine, 740.

Nitrates and nitrites, formation and
destruction of, in artificial solutions,
and in riverand well waters, 632.

Nitric acid, action of phosphorous oxy-
chloride on, 224.

——— peroxide, action of chlorine on,
226.

Nitrification and denitrification, alter-
nate, 669.

-—— influence of organic matter on, 667.

is organic carbon essential to? 651.

of ammonium and potassium thio-
cyanates by soil, 637.

—— ofammonium salts, 6438. 654.

of ethylamine by soil, 633.

of gelatin by soil, 641.

of urea, 639.

— of urine, 642,

Nitrifying organism, period of incuba-
tion of, 679.

Nitroxyl chloride, non-existence of, 222.

Nitryl chloride, non-existence of, 222.

Oak, scarlet, analyses of white and
green Jeaves of, 839.

Obituary notices, 342.

Oxygen, action of hydrogen on, 107.

action of, on hydrochloric acid,

under the influence of light, 608.

forms of, obtained in the electro-

lysis of aqueous sulphuric acid, 591.

free, estimation of, in water, 751.

Oxyketones, formation of, 54.

Ozone, formation of, in the electrolysis
of aqueous sulphuric acid, 591.

Parasites, vegetable, action of ferrous
sulphate on, 119,

Perezonoxime, 721.

Perkin’s reaction, 317.

874 INDEX OF SUBJECTS.

Phenanthro-eurhodine, 400. I

Phenol constituents of blast-furnace tar
obtained by the Alexander and
McCosh process at the Gartsherrie
~ rrr examination of. Part I,

Phenylbromisosuccinic acid, p- and o-
nitro-, 362, 363.

Phenylbutindicarboxylic acid, 366.

Phenyldibromisosuccinic acid, 360.

— m-nitro-, 361.

Phenylhydrazine cyanuric chloride, 742.

Phenylpropionic acid, 441.

Phenylsulphonic anhydride, 692.

Phloretin, 860.

Phosphorus chorobromides, 815.

oxychloride, action of nitric acid

on, 224.

suboxide, 834.

—— tetroxide, 833. i

—— trichloride, action of bromine on, I
815.

Phthalic acid, 8-chloro-, 526.

B-sulpho-, 510.

— acids, monobromo-, 187.

amide, 8-sulpho-, 521.

—— anhydride, chloro-, 528.

B-sulpho-, 515.

chloride, chloro-, 527.

—— chlorides, 8-sulpho-, 520, 521.

Phthalimide, 8-chloro-, 529.

Pipitzahoic acid, 709, 712.

action of hydroxylamine and
phenylhydrazine on, 723, 724.

— — amido-, 720.

anilido-, 717.

—— —— dibromide, 731.

o- and p-toluido-, 718, 719.

Potassium copper sulphate, anhydrous,
heat of dissolution of, 300

anhydrous, specific

heats of three modifications of, 14.

—- hexahydrated, heat of
dissolution of, 301.

modifications and heats

of dissolution of, 1.

magnesium sulphate, anhydrous,

heat of dissolution of, 297.

hexahydrated, heat of

dissolution of, 298.

modifications and heat

of dissolution of, 7.

nitrate, action of sulphuric chlor-
hydrin on, 225.

— —— action of sulphury] dichloride
on, 226.

sulphate, heat of dissolution of, 306.

—— thiocyanate nitrification of, by
soil, 637.

Propione, preparation of, by Perkin’s

method, 323.

——

INDEX OF SUBJECTS.

Propionic acid, hydrated, magnetic
rotation of, 780.
vapour-pressures of, 766, 774,

776.

Propyl alcohol, action of Bacterium
aceti on, 177.

normal, vapour-pressures of,
763, 771, 773.

a-Propyl-8-chlorocinnamic acid, 163.

Quercus rubra, analyses of white and
green leaves of, 839.

Raffinose, presence of, in barley, 70.

River waters, destruction and formation
of nitrates in, 632, 656, 664.

Rose trees, experiments with ferrous
sulphate on, 122.

Salicylic aldehyde, action of, on malonic
acid, 365

Salts, double, heat of formation of, 287.

Selenites, relation of,.to sulphites, 584.

Selenium sulphoxide, 583.

Silicon, condition of, in pig iron, 215.

influence of, on the properties of
cast iron, 130.

Smilax glycyphylla, sweet principle of,
857.

Sodium «-naphthol -p -diazodiphenylsul-
phonate, 383.

—— £-naphthol - p - diazodiphenylsul-
phonate, 383.

—— phenol-p-diazodiphenylsulphonate,
382.

—— quinol-p-diazodiphenylsulphonate,
382.

— resorcinol - p - diazodiphenylsul-
phonate, 382.

—— sulphate, anhydrous, heat of dis-
solution of, 302.

decahydrated, heat of dis-
solution of, 303.

Steam and carbonic oxide, action of
induction sparks on, 103.

Sugar, cane-, action of Bacterium aceti
on, 181.

Sugars of some cereals and of germinated
grain, 58.

Sulphates, double, modifications of,
Parts I and II; Part II. Specific heat
determinations 1, 12.

875

Sulphates, multiple, 123.

Sulphites, action of phosphorus oxy-
chloride on, 588.

constitution of, 533.

a-Sulphophthalic acid, 512.

8-Sulphophthalic acid, 510.

anhydride, 515.

—— amide, 521.

chlorides, 520, 521.

Sulphur sulphoxide, 583.

Sulphuric acid, electrolysis of aqueous,
with reference to the forms of oxygen
obtained, 591.

— chlorhydrin, action of potassium
nitrate on, 225.

—— oxide, note on, 584.

Sulphuryl dichloride, action of potas-
sium nitrate on, 226.

Taurine, derivatives of, 486.

Tellurium sulphoxide, 583.

Temperature, influence of, on the heat
of chemical combination, 260.

Terpenes, refraction and dispersion
equivalents of, 612.

Tetramethylammonium bromide, action
of iodine monochloride on, 850.

and chloride, action of bro-

mine and chlorine on, 852.

— action of iodine on, 847.

— dibromiodide, 848.

— — action of ammonia on, 849.

dichloriodide, 850.

—— iodide, action of bromine on, 848.

—— — action of chlorine on, 849.

—— salts, action of halogens on, 846,

— sulphate, action of bromine and
chlorine on, 853.

Tetrethylphosphonium iodide, action of
bromine on, 854.

Thiocarbamide, derivatives of, 190,

—— non-nitrifiability of, 639.

Thymoquinone, amido-, 725.

Toluidylmelamine, 742.

Toluylenediamidocyanuric chloride, 741.

o-Toluylenediamine, 259.

m-Tolylene diisocyanate, 257.

Trimellitic acid, 531.

Trimethyldiethylamidobenzene, 813.

Trimethyltaurine, decomposition by
alkali, 486.

Tri-a-napthylmelamine, 315.

Undecylenic acid, constitution of, 205.
magnetic rotation of, 206.

876

Urea, action of hexabromacetone on, 693,
743.

— nitrification of, 639.

Urine, nitrification of, 642.

Vanadious sulphate, electrolytic pre-
paration of, 822.

Vanadium compounds, some new, 30.

Vapour-densities of chloral ethyl-
alcoholate, 685.

pressures of alcohols and organic

acids, 761.

of alcohols and organic acids,

relations between, 761.

of bromine and iodine, 453.

of mercury, 37.

—

Vegetable albinism, a chemical study of.
IIL. Experiments with Quercus rubra,
839.

Vinegar plant, nature of, 433.

Vinylethylene, its
oxidation, 81.

preparation and

INDEX OF SUBJECTS,

Water, estimation of free oxygen in,
751.

—— of crystallisation, 280, 411, 690.

observations on, 788.

Waters, river and well, destruction and
formation of nitrates in, 632, 658, 664.

Well waters, destruction and formation
of nitrates in, 632, 658, 664.

Xx.
m-Xylenol, 23.

Zirconium, method of separation and
estimation of, 481, 149.
pentoxide, 483, 149.

ERRATA IN VOL. XLIX.

Page Line

/7 7 from top, for “15633” read “ 16233.”
” 8 » » » “9138” ,, “9738.”
‘ 9 4, ss » “(91388-9709 =) — 571” read “ (9738—9709=)29.”
» 12618 , 4, 4, “2649” and “731” read “3249” and “1331.”

» 15—20 , 4, #5 “mone of them .. . . would do” read “the
only one, the a-modification, which dissolves with
the heat evolution with which the mixed salts
would, is blue, whereas the mixture would be white,
and that, of the two other modifications, one dis-
solves with an evolution of 3249 cal. less, and the
other 1331 cal. more, than a simple mixture would.”

15 14, 9» » “—571” read “29.”
481 18 ” ” ” ™ P,S,,” ” 5 P,S;.”

535 6 4, 5» 9» “* mercurosic,” read “ mercuric.”

557 10 ,, 5» 4, “ mercury sulphite,” read “ mercurosic sulphite.”
558 9 ,, 9» » ‘ Sartorious,” read “ Sartorius.”

560 8

as ~- from = for “ sulphate,” read ‘‘ sulphite.”
5681 16, ,,

563 17. ,, 5, for “ no action,” read “ an action.”
572 8 ,, bottom, for “ sulphide,” read “ sulphite.”

~ “ ‘SO. -O ) “ 0:8 k ”
575 «12s, ‘top, for He<0 so. o> Hee" read Hes<Os0.0> Hee
583 8 4, 5: 9» “make,” read “makes.”

583 24 5, 9»: 9 “* H.SOs,” read “ 2H,SO;.”
587 13 ,, bottom, for “ Ag0s, 30, ” read “ Ag0s, 50,
3 Hg- ? Ag- 2

625 18 from bottom, for “ NH-Y-NR,” read “ NH,-Y-NHR.”

- S@ « . » “NH,-Y-NR,” , “ NH,-Y-NHR.”

a 14, » » “NRY.NR,” ,, “NHR-Y-NHR,” and

» “X°NHR,,” , “X-NHR.”

747 21 ,, top » “although benzene,” read “although azoxybenzene.”

757 6 ,, bottom ,, “138 c.c.,” » “10 cc.”
710 17 ,, top » “ extracted,” » “estimated.”

HARRISON AND SONS, PRINTERS IN ORDINARY TO HER MAJESTY, ST. MARTIN'S LANE,

No. COLXXXVIIL]

[Novemper, 1886. |} m

JOURNAL

THE CHEMICAL SOCIETY.

CONTAINING

THE PAPERS READ BEFORE THE SOCIETY,

4nD

ABSTRACTS OF CHEMICAL PAPERS PUBLISHED IN OTHER JOURNALS.

Committee of Publicution:

H. E. Arnmstronea, Ph.D., F.R.S.

8S. U. Proxrrine, M.A.

F. R. Jape, M.A., Ph.D., F.R.S. | BR. T. Prmeeron, Ph.D. :
R. Mextpota, F.RB.S. | W. Jd. Rossent, Ph.D., F-RS.
Hueco Méuumr, Ph.D., F.R.S. J. Mrittar Txomson, F.R.S.E.
F. J. M. Pager, B.Sc. | T. E. Tuorps, Ph.D., F.B.S. “4
W. H. Psze«u, Ph.D., F.B.S.
Editor:
C. E. Grovrzs, F.R.S.
Sub-editor :
A. J. GREENAWAY.

Sbstractors ;
H. Baker. | N.H.J. Mruuzr, Ph.D.
D. Beyprx. | G. H. Morris, Ph.D.
C. H. Borwamuery. J. M. H. Muwno, D.Sc.
B. H. Broves. A. Purp.
©. F. Cross. HE. W. Prevost, Ph.D.
J. K. Crow, D.Sc. R. Rovriepes, B.Sc.
Wrnpuam R. Dunstan. M. J. Saurer.
JOSEPH FLETOHER. JaMES TaYLor, BS.
P. F. Franxptann, Ph.D., B.Sc. A. THILLoT.
W. D. Hatirvurron, M.D., B.Sc. | L. T. THorwz, Ph.D.
J. P. Laws. | V. H. Vurey, "MLA.
D. A. Lous. | EP. Wuirs.
T. Maxwett, M.D., B.Sc. | W. C. Wiinrams, B.Sc,

LONDON:

J. VAN VOORST, 1, PATERNOSTER ROW.

HARRISON AND SONS, PRINTERS IN ORDINARY TO HER MAJESTY, ST. MARTIN'S LANE,

ISSUE OF JOURNAL TO SUBSCRIBERS.

The Journal of the Chemical Society is regularly despatched, post
free, by the Publisher (J. Van Voorst, 1, Paternoster Row, E.C.) to
subseribers—English and Foreign, within the Postal Union—who
have paid to him in advance the annual subscription of 30s.

On payment of one shilling annually, post-cards with the
titles of the papers to be read will be sent to Fellows before
each meeting of the Society.

Fellows who wish to have these cards are requested to
_. forward their names and subscriptions to Messrs. Harrison

- and Sons, 45, St. Martin’s Lane, London, W.C.

CHEMICAL SOCIETY RESEARCH FUND.

AMOUNT RECEIVED SINCE JULY Isr, 1885.

The Worshipful the Drapers’ Company ..

: Chemical Laboratory, W. J. RUSSELL Treasurer.
" $t, Bartholomew's Hospital, E.C.,
. July, 1385.

@Olficers and Council,

Elected March 30, 1886.

PRESIDENT.)
H. MULLER, Px.D., F.R.S,

VICE-PRESIDENTS

WHO HAVE FILLED THE OFFICE OF PRESIDENT,

Str F. A. ABEL, O.B., D.C.L. F.R.S.
WARREN DE LA RUE, D.O.L., F.B.S.
E. FRANKLAND, D.O.L., F.R.S.

J. H. GILBERT, Pu.D., F.R.S.

J. H. GLADSTONE, Pua.D., F.B.S.
A. W. HOFMANN, D.O.L., F.B.S.
W. ODLING, M.B., F.R.S.

W. H. PERKIN, Pua.D., F.B.S.

Sm LYON PLAYFAIR, Pa.D., K.0.B.,
F.R.S.

Srr HENRY E. ROSCOK, LL.D.,
F.R.S.

A. W. WILLIAMSON, LL.D., F.B.8.

VICE-PRESIDENTS.

W. CROOKES, F.RB.S.
J. DEWAR, M.A., F.RS.
D. HOWARD.

G. D. LIVEING, M.A, F.R.S.
T. E. THORPE, Pu.D., F.B.8,

| W. A. TILDEN, D.Sc., F.R.S.

SECRETARIES.

H. E. ARMSTRONG, Pu.D., F.B.S.

| J. M. THOMSON, F.RB.S.E.

FOREIGN SECRETARY.
F. R. JAPP, M.A., PuD., FBS.

TREASURER.
W. J. RUSSELL, Pu.D., F.R.S.

OTHER MEMBERS

H. T. BROWN.

T. CARNELLEY, D.Sc.
M. CARTEIGHE.

Ff. CLOWES, D.Sc.

A. E. FLETCHER.

R. J. FRISWELL.

OF THE COUNCIL.

R. MELDOLA, F.R.S.

R. MESSEL, Pua.D.

J. A. BR. NEWLANDS.
S. U. PICKERING, M.A.
W. RAMSAY, Ph.D.

T. STEVENSON, M.D.

‘ '

IMPORTANT NOTICES.

ABSTRACTS OF THE PROCEEDINGS OF THE
CHEMICAL SOCIETY.

The attention of Subscribers to the Journal of the Chemical

} Society is directed to the fact that abstracts of the papers read

at the bi-monthly meetings and the discussions thereon are
now published by the Society, being usually issued to Fellows
about a week after the meeting. The Council have resolved
to place these Abstracts of Proceedings on sale from the com-
mencement of the new year at a price of seven shillings and
sixpence per annum, post free. Intending subscribers are
requested to apply to Mr. VAN VOORST, PATERNOSTER
ROW, LONDON.

NOTICE TO AUTHORS OF PAPERS.

To facilitate the preparation of the abstract accounts of the
Proceedings at the meetings of the Society, which are now

4 published, all authors are requested to furnish abstracts of

their communications, and to send their papers so that they
may be in the hands of thé~Secretaries, if possible, on the
Monday before the day of meeting.

PROCEEDINGS OF THE ROYAL SOCIETY.

Subscriptions of 10s. can now be received by Mr. Hall for the cur-
rent volume 41 of the Royal Society’s Proceedings; also for volumes
27, 30, 37, 38, 39, and 40.

_All communications for the Journal of the Society should be
addressed to ‘“‘ The Secretaries” at Burlington House.

Anthors are requested to send their corrected proofs without delay
tothe Editor, Kennington Green, 8.E.

Abstractors are requested to send their MSS. and corrected proofs
to the Sub-editor, Frognal, Hampstead, N.W.; and communicate
hange of address to the Priaters, as well as to the Editors.

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Texts

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Journal of the Chemical Society. Transactions 1886: Vol 49"

See other formats